(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Effectiveness of Alternative Skid Reduction Measures : Final Report, Volume 4 - Criteria For Improvement of Pavement Surface Macrotexture"

No. FHWA-RD- 79-25 



.A3 

[;il!'" XTIVENESS OF ALTERNATIVE 
^Uaii REDUCTION MEASURES 

Vol. IV. Criteria for Improvement of Pavement 
Surface Macrotexture 

November 1978 

Final Report 



^.-^s*''■'' 



DEPARTMF.^ 
TRANSPOBTAi 



APR 2 \o.^ I 

'■• LIBRARY I 




aSL5£2%>. 




•Sr/iTEs o* 



Document is available to the public through 
the National Technical Information Service, 
Springfield, Virginia 22161 



ij'' 



Prepared for 

FEDERAL HIGHWAY ADMINISTRATION 
Offices of Research & Development 
Environmental Division 

Washington, D.C. 20590 



FOREK^ORD 



This report is part of a final report consisting of an executive summary 
and four volumes. The executive summary provides a synopsis of the 
research. Volume I describes the evaluation of accident rate-skid number 
relationships; Volume II describes the development of the benefit-cost 
model; Volume III presents the computerized benefit-cost model and 
instructions for its use; and Volume IV summarizes methods of measuring 
and achieving macrotexture. It will interest those concerned with 
pavement surface characteristics and the selection of accident reduction 
measures. 

This research is included in Project IH, "Skid Accident Reduction" of 
the Federally Coordinated Program of Research and Development. 
Mr. George B. Pilkington II is the Project Manager and Mr. Philip Brinkman 
is the Task Manager. 

One copy of this report is being distributed to each FHWA regional office. 




Director, Office of Research 
Federal Highway Administration 



NOTICE 

This document Is disseminated under the sponsorship of the Department of 
Transportation in the interest of information exchange. The United States 
Government assumes no liability for its contents or use thereof. The 
contents of this report reflect the views of the contractor, who is 
responsible for the accuracy of the data presented herein. The contents 
do not necessarily reflect the official views or policy of the Department 
of Transportation. This report does not constitute a standard, specification, 
or regulation. 

The United States Government does not endorse products or manufacturers. 
Trade or manufacturers' names appear herein only because they are considered 
essential to the object of this document. 



I. Report No. 

FHWA-RD-T9-25 



2. CavontMont Acci 



No. 



.■4. Till* ond Swblitio 

Effectiveness of Alternative Skid Reduction 
Measures. Volume IV: Criteria for Improve- 
ment of Pavement Surface Macrotexture 



7. A«ri«H^.) Q^ y^ Harwood, R. R. Blackburn and 
P. T. Hppnan 



3. R«cipi«nt°s Cai 1I09 No 



DEPARTMENT OF 
TRANSPORTATION 



APR 2 1980 

5. R.po,.Oor. ^- LIBRARY 

November 11978 



6. Performing Organization (la<l« 



8. Parfemtng Orgona lotion Report No. 



3824-D 



9. Per^rming Orgoniiotion Nomo and Address 

Midwest Research Institute 

425 Volker Boulevard 

Kansas City, Missouri 64110 



10. Wori Unit No. (TRAIS) 

31H5-014 



1 1 ■ Controct or G'ont No. 

DOT-FH-11-8120 



12. Sponaaring Ag en cy Nome and Address 

U.S. Department of Transportation 
Federal Highway Administration 
Offices of Research and Development 

Washington, D.C. 20590 



13. Type of Report and Period Covered 

Final Report 

June 1973-November 1978 



I-4- Sponsoring A9«ncy Co^o 



IS. SuppleoMntory Notes 



FHWA Contract Manager: P. Brinkman, HRS-it3 



V6. Abstract 

This voliime presents a guide to the role of pavement macrotexture in 
pavement skid resistance and accidents* An overview of the fundamentals 
of tire-pavement skid resistance is presented. The methods of measuring 
macrotexture in the field and in the laboratory are described based on a 
review of relevant literature* The measurement methods currently employed 
by state highway departments are identified; the sand patch method is the 
most widely used and accepted method in the United States for measuring 
pavement macrotexture* 

Methods of providing macrotexture in new pavements and restoring macro- 
texture in existing pavements are described* The methods of providing 
macrotexture in new pavement include open-graded asphalt surface courses 
and texturing of portland cement concrete surfaces; the methods of restor- 
ing macrotexture to existing surfaces include open-graded asphalt overlays, 
pavement grooving, cold milling, and seal coats. A cost-effectiveness 
analysis procedure for alternative methods for improving pavement macro- 
texture is presented* Such analyses can be used as the basis for cost- 
effective warrants for pavement macrotexture improvements* The develop- 
ment of cost-effective warrants is illustrated by a nijmerical example* 



t7. Key Words 

Macrotexture 
Pavement Texture 
Skid resistance 



Texture measurement 
Cost-effectiveness 



18. Distribution StoteMent 

Document is available to the public 
through the National Technical 
Information Service (NTIS), Springfield, 
Virginia 22l6l. There is a charge for 
CQ-pies ordered from NTIS. 



19. Saevnty Claasil. («< i4ita report) 

Unclassified 



3D. Soeynty Clossrf. (o' >Ms poge) 

Unclassified 



21< No. al Pages 

144 



22. Prjeo 



PREFACE 



This is volume four of a four -volume set prepared by Midwest Re- 
search Institute for the Federal Highway Administration under Contract No. 
DOT-FH-8120. Mr. Henry C. Huckins of the Implementation Division of the 
FHWA Office of Development served as Contract Manager for the preparation of 
this volume. The project also benefited from the comments and suggestions 
of staff members of the FHWA Office of Research including Mr. Charles P. 
Brinkman, Mr. George Pilkington, and Mr. Burton Stephens. 



We also wish to acknowledge the contributions of 18 state high- 
way and transportation agencies who cooperated with the project. The co- 
operating agencies are the California Department of Transportation, the 
Connecticut Department of Transportation, the Kansas Department of Trans- 
portation, the Louisiana Department of Highways, the Maine Department of 
Transportation, the Maryland State Highway Administration, the Massachusetts 
Department of Public Works, the Michigan Department of State Highways, the 
Mississippi State Highway Department, the North Carolina Department of Trans- 
portation, the Ohio Department of Transportation, the Pennsylvania Depart- 
ment of Transportation, the Rhode Island Department of Transportation, the 
South Carolina State Highway Department, the Texas Department of State 
Highways and Public Transportation, the Washington State Highway Commission 
and the West Virginia Department of Highways. 

The work reported in this volume was carried out in the Economics 
and Management Science Division under the administrative direction of 
Dr. William D. Glauz. Mr. Robert R. Blackburn, Manager, Driver and Environ- 
ment Programs, was the Principal Investigator for this effort. Mr. Blackburn, 
together with Mr. Douglas W. Harwood, Associate Traffic Engineer and 
Mr. Patrick J. Heenan, Junior Engineer, were co-authors of this volume. 
Mr. Jerry L. Graham, Associate Traffic Engineer, also contributed to the 
project. 



Approved for: 

MIDWEST RESEARCH INSTITUTE 




A. E. Vandegrift, Direc4^t 
Economics and Management 
Science Division 




11 



TABLE OF GO^^^ENTS 



I* Introduction* •••••••••••••••••••••••• 1 

II* Fundamentals of Tire-Pavement Skid Resistance •••••••• 3 

A* Over\7lew •••••••••••••••••••••• 3 

B» Role of Pavement Surface Properties in Skid 

Resistance •••••••••••••••••••• 8 

III* Pavement Macrotexture Measurement ••••••••••••••12 

A* Overview of Measurement Methods* ••••••••••12 

B* Current State Practice for Macrotexture 

Measurement* **••••••••••••••••• 38 

IV* Methods of Achieving Macrotexture in New Pavements* • • • • • 43 

A^ Open-Graded Surfaces for New Bituminous 

Pavements* **•••••••••••••••••• 43 

B^ Texturing of Portland Cement Concrete Surfaces • • • 51 

¥• Methods of Restoring Macrotexture for Existing Pavements • • • 58 

A^ Open-Graded Asphalt Overlays ••••••••••••58 

B* Pavement Grooving* •••••••••••••••••60 

C* Cold Milling •*• • 62 

0* Seal Coats •*•••••••••••••• 65 

VI* Cost-Effectiveness of Pavement Macrotexture Improvements* • • 69 

A* General Approach *•******•••••••••• 70 

B* Construction Cost Estimates* ••*•••*••••• 71 

C* Accident Reduction Effectiveness ••••••••••73 

D* Benefit-Cost Comparisons ***•*••••••••• 79 

E* Interpretation of Benefit-Cost Ratios* •**••*• 93 

VII* Warrants for High-Macro texture Surface Courses* ••96 

A* Current State Practices* •**•••******•• 96 
B* Development of Cost-Eff ective Warrants for Use of 

High-Macro texture Surface Courses* ••••••••98 

VIII* Conclusions **•**•*•••*•••**•******* •103 



iii 



TABLE OF CONTENTS (continued) 

Page 

IX* Recoinmendations* ••••••••••••••••••••• 105 

X* References •••••••••••••••••••••••• 106 

Appendix A - FHWA Design Procedure for Open-Graded Asphalt Mixtures* 114 

Appendix B - Typical Specifications for Constructing and Measuring 

Pavement Grooves* *•*****•***•**•••• 133 

List of Figures 

Figure Title Page 

1 Relation Between Friction Demand and Skid Resistance • • • 4 

2 Skid Number-Speed Relationships for a Variety of Pavement 

Surfaces •••*•••*•••••••••••••••• 6 

3 Graph Showing Approximate Correlation Between Average 

Pavement Texture Depth and Outflow Time* ******** 22 

4 Skid Resistance Change Frcxn Traffic on PGC Pavements * * • 55 

5 Example of Warrants for Open-Graded Asphalt Overlays • • • 100 

6 Example of Warrants for Open-Graded Asphalt Overlays • • • 101 

7 Chart for Determining Surface Constant (Kc) of Coarse 

Aggregate* A********************** 118 

8 FHWA Vibratory Compaction Apparatus* *********** 120 

9 Tamper Foot and Extension* **************** 121 

10 Cylindrical Mold for Testing Granular Materials* • * • * • 122 

11 Determination of Optimum Fine Aggregate Content* ***** 126 



IV 



TABLE OF CONTENT S (concluded) 

List of Tables 

Table Title Page 

1 Test Results for Dense-Graded and Open-Graded Asphalt 

Pavement Surfaces* •••••••••••••••••• 49 

2 Open-Graded Asphalt Friction Courses for New Pavements • 50 

3 Initial Texture Depths for Various Finishes. •••••• 51 

4 Texture Depth Degradation After 30 Months. .••*••• 54 

5 Texturing of New Portland Cement Concrete Surfaces • • • 57 

6 Pavement Grooving. .•••••••••••••••••• 63 

7 Retexturing Pavements by Cold Milling. •.••••••• 66 

8 Estimates of Initial Microtexture, Macrotexture and 

Skid Number Used in the Cost-Effectiveness 

Analysis •••••••••••..•..•...... 74 

9 Example of Determining Equivalent Uniform Annual 

Reduction in Accident Rate (missing) •••••••• 78 

10-17 Benefit-Cost Calculation (missing) for a 1,6-Km etc. . . 81 

18 Summary of Unadjusted Benefit-Cost Ratios for Pavement 

Macrotexture Improvements. «•••.••.•..... 90 

19 Wet-Pavement Exposure Time Adjustment Factor ..*•.. 91 

20 Traction Demand Adjustment Factor. ....•..•••• 92 



(A (A 

£ J= M> -CI O 

w o O I- — 

£ .£ 5 5. E 



S S S £ 
ff V cr o 
tn w) n « 



s 



.° Z -• 



M 

PC 
O 
H- 
U 



M 



w O <D ^ 

CO (S n M 
a 3 5 u 

sr sr yj 



u E .3 



2 .■= g 



a"* 



:sse||| 



s^ss « 


3 


*SJ 


ON ^o«^ 


< 


i5? 

o> a 



i i 



.r .s 3 3 



£ E E 



nil 


nil 


nil 


III! 


nil 


nil 


oz 

nil 


nil 


SI 

nil 


nil 


81 

nil 


nil 


nil 


nil 


91 

III! 


nil 


SI 

III! 


nil 


III! 


nil 


ei 

nil 


nil 


ZI 

nil 


nil 


II 

nil 


nil 


01 

nil 


nil 


6 

iin 


nil 


8 

nil 


ini 


nil 


III! 


9 

ini 


III! 


s 

nil 


iin 


♦ 

III! 


nil 


£ 

nil 


nil 


z 

nil 


nil 


I 


UI3 

nil 






9 


'I'T 


8 


TT 


TT 

7 


T T 


TT 

6 


TT 


TT 

5 


T T 


T T 

4 


TT 


TT 

3 


T T 


TT 

2 


TT 


TT 

1 


T T 

inches 



e 
e 
u 



" » s ■= 
£ 5 ?. E 



« 5 S (5 3 
« a ff S^ ^ 



113 O flO to ^ 
to O O N O 



•5 S 1 ' 



tr y tr 

new 



j_ 2 « 
E g" S 
(0 — e 



EEE E E 



E E E = .= -i X 



lAiooooonoo 



II 



S- 



|3 



« 2 -a 

£ S ^ u 'a 5- 



t g s a 

« — A ^ 






s-B 



o . 

e o 



o *» 

O 3 

U. to 



3? 

S a 



VI 



I. INTRODUCTION 



This volume, "Criteria for the Inqjrovement of Pavement Macro tex- 
ture," is the fourth volume of a four-volume set prepared as the result of 
a multiyear study of wet-pavement accidents. The project grew out of the 
increasing national concern over the hazards connected with driving on wet 
pavements. Under such conditions, the tire-pavement friction level is re- 
duced — often dramatically. If vehicle maneuvers demand higher friction 
levels, skidding will occur, potentially leading to an accident. The risk 
of such accidents tends to increase with higher traffic volumes and higher 
speeds. The traffic safety engineer must therefore develop and implement 
countermeasures that will either increase the tire-pavement friction level 
or reduce the demand for friction. 

The contract dealt with two aspects of this total problem. The 
first was to develop the relationships between pavement skid number and 
wet-pavement accidents for a variety of highway and traffic conditions. 
This work is documented in Volume I of the set. 

The second phase of the study involved the definition of a range 
of alternative solutions to the problems of maintaining the frictional re- 
quirements of drivers during wet-pavement conditions, and the development 
of a formal means to evaluate the cost-effectiveness of the solutions. 
This was accomplished through a conqjrehensive, computerized, benefit-cost 
model. The model incorporates, in detail, the relationships between ac- 
cidents and skid number and an indirect relationship developed between ac- 
cidents and pavement texture, so that the effects on accidents of changes 
in skid number or pavement texture can be predicted. It also treats ex- 
plicitly, via self-contained tabular data or user-supplied data, the ef- 
fects of geometric and traffic control countermeasures on a variety of ac- 
cident types. The details of the model and its development are given in 
Volume II. 

The third volume of the report, the User's Manual, presents im- 
plementation procedures. It shows how the user can apply the benefit-cost 
model, not only to wet-pavement accident reduction, but to an extremely 
broad class of benefit-cost evaluations of accident countermeasures. The 
manual is designed for two types of users: (1) highway engineers or ad- 
ministrators who need to understand the overall aspects of the model to 
use it effectively in decisionmaking; and (2) engineers and other persons 
who need to prepare data and actually use the model. 

Volume IV is a guide for highway engineers and administrators 
concerned with the improvement of pavement surface texture as a counter- 
measure for skidding accidents. It discusses the importance of pavement 
macrotexture in reducing skidding accidents. The volume then describes 



the methods of measuring pavement macrotexture and techniques for provid- 
ing macro texture in new pavements and restoring macrotexture in existing 
surfaces. A simplified version of the benefit-cost model described in 
Volumes II and III is used to perform a cost-effectiveness analysis of al- 
ternative techniques for improving pavement macrotexture. The criteria 
presented for the improvement of pavement surface macrotexture can be di- 
rectly implemented by state highway departments. Additional criteria, ap- 
plicable to the local climate and paving materials in specific states or 
regions of the country, can be developed using the general analysis ap- 
proach presented in this volume. 



II. FUNDAMENTALS OF TIRE-PAVEMENT SKID RESISTANCE 



This section provides a summary of the fundamentals of tire-pavement 
skid resistance. The discussion focuses on the contribution of pavement sur- 
face properties to the friction available at the tire-pavement interface. The 
information presented here is essential background for a complete understand- 
ing of the remainder of this report. 



A. Overview 

The potential for a skidding accident depends mainly on the speed 
of the vehicle, the cornering path, the magnitude of acceleration or braking, 
the condition of the vehicle tires, and the characteristics of the pavemient 
surface. On wet pavements, speed is the most significant parameter, not only 
because the frictional demand increases with the square of the speed, but also 
because the skid resistance at the tire-pavement interface decreases with in- 
creasing speed. Figure 1 is a generalization of these relationships showing, 
for a given degree of cornering or magnitude of acceleration or braking, how 
the margin of safety (difference between available skid resistance and fric- 
tion demand) decreases rapidly with increasing speed until a skid is imminent. 

Skid resistance is a general term used to describe the level of 
friction between a roadway surface and a vehicle tire. The tire-pavement skid 
resistance can be measured in several testing modes--locked-wheel braking, 
brake slip, drive slip, and cornering slip. A few states, primarily in the 
Western United States, operate cornering slip testers, also known as mu- 
meters. However, the locked-wheel braking mode has gained by far the widest 
acceptance for skid resistance testing throughout the United States. 

The most common measure of skid resistance in the locked-wheel brak~ 
ing mode is the skid number (3N), which is defined as 100 times the coefficient 
of friction determined with a locked-v^eel skid test. The standard procedure 
for skid number measurements in the United States is established by A3TM Stan- 
dard E-274-77. Testing is accomplished by a two-wheel trailer towed by a 
truck at a standard speed of 64.4 km/hr (40 mph). A layer of water with a 
standard thickness of 0,51 mm (0.02 in.) is placed on the pavement by nozzles 
located just in front of the trailer ^^eels. This requires a flow rate of 
about 0,60 liters/min/ram (4.0 gal./min/in.) of wetted uadth at 64.4 km/hr 
(40 mph)« The trailer brakes are activated to lock one or both of the trailer 
wheels on the wetted surface. A standard test tire, described on ASTM Speci- 
fication E-501 and mounted on a suitable 15 by 6-in» rim, is used. A trace 
of the wheel torque during braking is made and interpreted either manually 
or electronically to obtain the frictional force and the resultant skid nim- 
ber. 



Speed of Impending Skid 



c 

(U 

E 

> 

o 

a. 

I 




Tire- Pavement 
Skid Resistance 



Friction Demand of Vehicle 



Vehicle Speed- 



Figure 1 - Relation Between Friction Demand 
and Skid Resistance 



skid numbers have been measured for many pavements, and many states 
conduct a periodic skid number inventory of all state highways. Despite its 
ready availability, the skid number should not be the only factor on which 
implementation of skidding accident countermeasures is based. The skid num- 
ber is measured for one particular tire design and one particular water 
depth, whereas a variety of tire designs and water depths are encountered in 
the real highway environment. Also, the skid number by itself is a measure 
of available friction only and does not reflect the level of friction demand 
for the maneuvers that are required of drivers at any particular site. Fin- 
ally the skid number measured at one particular speed, 64.6 km/hr (40 mph), is 
not adequate to completely define the frictional properties of a pavement surface. 

The skid resistance curve in Figure 1 illustrates that the skid 
number of a pavement depends on the vehicle speed. Two pavements can have 
similar skid numbers at 64.4 km/hr (40 mph), but very different skid numbers at 
a higher speed that is more typical of operating conditions on rural highways. 
Figure 2 is an example of the skid number-speed relationships for a variety 
of pavement surfaces. Each of the pavements in Figure 2 has a nonlinear skid 
number-speed relationship, but these relationships are frequently approxi- 
mated by a straight line. The slope of this line is represented by the skid 
number-speed gradient. It is conventional to define the skid number-speed 
gradient as a positive number. This quantity can be determined from the re- 
sults of skid tests at two speeds as: 

SN - SN 
A B 

where SNG = Skid number-speed gradient between speeds A and 3 (SN/km/hr 

or SN/mph) 

2Na ~ Skid number at speed A, 

SNg = Skid number at speed B, 

V^ = Lower testing speed (mph or km/hr), and 

V-o = Higher testing speed (mph or kia/hr). 

In the recommended ASTM test method, speed A is 48.3 km/hr (30 mph) and speed 
B is 80.5 km/hr (50 mph). These speeds define a 32.2 km/hr (20 mph) interval 
centered on the standard test speed of 64.4 km/hr (40 mph). However, other 
combinations of speeds have also been used by the states to determine the skid 
number-speed gradient. 

The magnitude of the skid number-speed gradient indicates the rate 
at which skid number decreases with increasing speed. A pavement with a 
relatively flat skid number-speed relationship has a small skid number-speed 
gradient and will retain a higher skid resistance at increased speeds than 
a pavement \<n.th a relatively steep skid number-speed gradient. 




20 40 60 

Sliding Speed, V, mph 



80 



20 



40 60 80 

Sliding Speed, V, Km/Hr 



100 



120 



Figure 2 - Skid Number-Speed Relationships for a Variety of 
Pavement Surfaces-Lr./ 



It should be kept in mind that both the skid number and the skid 
number-speed gradient must be considered in evaluating a pavement. For ex- 
an^jle, Pavement 3 in Figure 2 has skid resistance qualities preferable to 
Pavement 4, despite Pavement 3's higher skid number-speed gradient, because 
it retains a higher skid number at speeds above 64.4 km/hr (40 mph)» 

The skid number of a pavement is known to change with tine due to 
the action of traffic passages. Different aggregates polish at differenc 
rates. Several states have developed wear and polish machines for testing 
the polish-susceptibility of aggregates in the laboratory. A discussion of 
wear and polish mechanisms that influence skid resistance can be found in 
Appendix B of Volume TI.86/ 

Although the skid number is the most commonly used measure of skid 
resistance, it has one important limitation that must be recognized by any- 
one who manages a wet-pavement accident prevention program based on skid num- 
ber improvements. Pavement skid numbers are known to vary v;ith the season of 
the year. The highest skid number values are generally observed in the spring 
months and the lowest values in the fall. Extreme seasonal variations as high 
as 30 have been observed, with more typical amplitudes in the range of 5 to 
15«2it' This phenomenon is presiimed to result from variations in ADT aggregate 
composition, temperature, precipitation, and other factors. The pattern of 
seasonal skid number variations is inconsistent. Gramling and Hopkins^' 
have assumed an "ideal" form of the seasonal variation curve that is roughly 
sinusoidal, but currently there is no reliable model for predicting these varia- 
tions. The Federal Highway Administration plans a major research effort, in 
cooperation with several state highway departments, to model and oredict 
seasonal variations in skid number. Until such information is available, 
engineers should exercise caution so that whenever possible skid number data 
used to justify improvement projects not overly biased by seasonal effects. 

Finally, it should be recognized that when the depth of water on a 
pavement is excessive, skidding accidents can occur even when the friction 
required by a particular maneuver does not exceed the pavement: skid number, 
measured with the ASTM Standard skid test. In this phenomenon, knoxjn as 
hydroplaning , the tire loses contact with the pavement and rides on a thin 
film of water. Very small accelerating, braking or cornering forces can 
initiate a skid in this condition. Therefore, at sites where climatic or 
geometric conditions are likely to produce excessive water depths on the 
pavement, consideration of factors other than skid number is required. 

An understanding of methods to increase high-speed skid-resistance 
and reduce the potential for hydroplaning requires consideration of the role 
of pavement surface properties in skid resistance, which is discussed in the 
next section. 



B. Role o£ Pavement Surface Properties in Skid Resistance 

When skid testing procedures and tire design are controlled, pave- 
ment surface properties are the only factors that (theoretically) influence 
skid resistance. An understanding of the role of pavement surface properties 
in skid resistance is particularly important to the highway engineer because 
he exercises some direct, continuing control over the condition of the high- 
way pavement, while tire design, vehicle characteristics, and driver behavior 
are at least one step removed from his control. 

The pavement surface properties that influence skid resistance can 
be divided into two categories: microtexture and macrotexture. Pavement 
microtexture consists of the microscopic asperities on the surface of indi- 
vidual pieces of aggregate, and to a lesser extent on the surface of the pave- 
ment binder (asphalt or portland cement). Microtexture is what makes a piece 
of aggregate feel smooth or rough to the touch. By contrast, macrotexture 
consists of the large scale asperities associated with voids in the pavement 
surface between pieces of aggregate. Thus, in the simplest terms, microtex- 
ture is determined by the aggregate surface and macrotexture is determined 
by the distribution of aggregate sizes and the manner in which the individual 
pieces of aggregate are assembled to form a pavement surface. Good skid re- 
sistance requires both good macrotexture and good microtexture. Separate 
discussions of macrotexture and microtexture are found below. 

1. Microtexture ; The low-speed skid number of a pavement surface 
is a function of the sliding resistance between the tire and the wetted as- 
perities of a pavement surface, where the tire contacts the pavement. This 
sliding resistance is increased if the pavement surface is rough or has dis- 
tinct asperities on a microscopic scale. Thus, low-speed skid number is 
closely related to microtexture. 

Several methods have been used to measure microtexture. A study.;:—' 
performed in Texas in 1970 reviewed several methods of microtexture measure- 
ment including subjective estimation and stylus tracing. The so-called Surf- 
indicator, a stylus tracing method, was subjected to limited field trials, 
but did not prove completely satisfactory. Dahir and Henry-i^^ used profile 
tracing to determine the microtexture of pavement surface directly. Spe- 
cific measures of microtexture considered by Dahir and Henry were root mean 
square texture height, arithmetic mean texture height, and root mean square 
texture slope. However, profile tracing to determine microtexture parameters 
is time consuming and not well suited to field measurements. 

The most frequently used method for measuring microtexture is the 
British Portable Tester. The British Portable Tester is a dynamic impact 
device used to measure the energy loss when a pendulum with a rubber slider 
contacts a test surface. A standard procedure for use of the British Port- 
able Tester is given by ASTM Standard E-303-74. The tester provides a read- 
ing called the British Portable Number (BPN) . The BPN of typical pavements 

8 



ranges from 55 to 90, with the larger values representing pavements with 
higher microtexture. Leu and Henryitl' have established a relationship be- 
tween microtexture (represented by BPN) and low-speed skid resistance 
(represented by the zero-speed intercept of skid number); 

SN = -31.0 + 1.38 BPN 
o 

where SN = Zero- intercept skid number, and 

BPN = British Portable Number, 

This relationship is based on data for 20 pavements and was established with 
a correlation coefficient of 0,75. The relationship indicates that pavement 
surfaces with good microtexture have good low-speed skid resistance. How- 
ever, the skid resistance at higher speeds is dependent on macrotexture as 
well as microtexture, 

2. Macrotexture t Pavement surface macrotexture has an extremely 
important role in the prevention of wet-pavement accidents. Good macrotex- 
ture provides a channel for water to escape from the tire-pavement inter- 
face. The ability of water to escape from the tire-pavement interface has 
a significant effect on the skid resistance of pavements at speeds above 
48.3 km/hr (30 raph). Two pavements with similar microtexture will differ in 
skid number at 64.4 km/hr (40 mph), if they differ in macrotexture. Thus, 
the theory of skid resistance suggests that macrotexture should be closely 
related to the skid number-speed gradient. Such a relationship has been 
demonstrated empirically in several studies, including the work of Schulze 
and Beckmann*!".' More recently, several researchers including Henry and 
Hegmon,2it' Veres, et al.,^' and Gallaway and Rose^' have found that the 
percent skid number gradient (PSNG), also known as the normalized skid num- 
ber gradient, and defined as: 



PSNG = - ^ —^ = ^^ (100) 




is more highly correlated with macrotexture than SNG. 

A number of methods have been used to measure pavement macrotexture 
including the sand patch, sand track, grease patch, outflow meter, profile 
tracing, stereophotographic, light stylus, laser and light depolarization 
methods. These and other methods are described in detail in section III of 
this report. The measures of macrotexture that have been used in studies 
of pavement surface texture include mean void width, average texture depth 
and root mean square texture height. 






Schulze and BeckmannZ^.' developed the following relationship be- 
tween skid number-speed gradient and mean void width: 



X = 0.876 - \J0.2376T - 0.04725 



where X = Skid number speed gradient between 20 km/hr (12.5 mph) and 60 
km/hr (37 ,5 mph), and 

T = Mean void width (mm) (1 ram = 0,04 in.). 

This relationship is based on data from 48 pavements and was established with 
a correlation coefficient of 0,87. 

Leu and Henry-^' have developed the following relationship between 
percent skid number speed gradient and macrotexture: 



PSNG = 23.04 (MD)"'^'^^ 



where MD = Average texture depth (mm) as determined by the sand patch 
method (1 mm = 0,04 in.). 

This relationship is based on data from 20 pavements and was established with 
a correlation coefficient of 0,96. 

3. Prediction of skid number from texture parameters : The predic- 
tion of skid number from texture parameters is both useful as a tool in the 
management of pavement surface improvement programs and provides insight into 
the roles of microtexture and macrotexture. The model developed recently by 
Dahir and Henry-L2' is presented here for illustrative purposes and is used in 
the cost-effectiveness analysis in Section VI of this report. This model is 
known as the Penn State model and it has the general form; 

C V 

where SN = Skid number at any speed V, 
V = Speed (km/hr), 
C_ = Zero speed intercept (correlated with microtexture), and 
C. = Function of macrotexture parameters only, 

10 



Expressions for Cj^ and C have been developed empirically by Leu and Henry 
from data for 20 pavement surfaces. 

The term Cq represents the skid resistance at low speeds and is, 
therefore, a function of microtexture alone. Cq is the zero-speed intercept 
of skid number, given above by Leu and Henry as: 

Cq = -31.0 + 1.38 BPN 

The macrotexture influence is expressed in terms of the percent skid number 
gradient. By differentiating the general prediction model, it can be shown 
that ; 

PSNG = -100 C 

Since, Leu and Henry have demonstrated that: 

PSNG = 23.0 (MD)"^*'^'^ 

Then, 

Ci = -0.230(MD)"°-^^ 

The relationships for C and C, , when substituted in the general 
model yield: 

SN^ = (-31.0 + l,38BPN)e-°'230V(MD)'" 

The coefficients of this model are based on an extremely small 
data set and are probably dependent on type of pavement surface. However, 
the model does illustrate separate influences of microtexture and macro- 
texture and demonstrates the sensitivity of skid number to each. 



11 



Ill, PAVEMENT MACROTEXTURE MEASUREMENT 



The qualitative influence of pavement surface properties, namely 
microtexture and macrotexture, upon the tire-pavement skid resistance has 
been known for a number of years. In recent years, there has been a need 
to quantify this influence. This need has spurred the interest of various 
agencies in the United States and abroad to engage in the development of 
methods for measuring the pavement surface texture, and in particular, pave- 
ment macrotexture. This section of the report discusses the general sub- 
ject area of pavement macrotexture measurement. Section A presents an 
overview of the most commonly or recently used macrotexture measurement 
methods. Where possible, it also presents a discussion of the precision 
and accuracy of the measurement techniques together with data on their 
known intercorrelations. The current practice of a selected sample of 
state highway departments in making macrotexture measurements is described 
in Section B. 



A. Overview of Measurement Methods 

A review of the literature revealed numerous methods for measur- 
ing pavement surface texture. Some of the techniques are in very limited 
use and their descriptions are found only in obscure literature. Descrip- 
tions of many of these methods have been reported by researchers including 
Rose, et al.,-^' Hegraon and Mizoguchi,— ' Dahir and Lentz,JiZ' Henry and 
Hegmon,~' Lees and Katekhda,~' Rose and Gallaway,— ' and Apostolos, et 
al.— ' Succinct descriptions of 28 of the most commonly or recently used 
methods of surface texture measurement are presented in this section. 
Descriptions of two additional methods that have not been used to measure 
pavement surface texture are also included because of their potential use 
as pavement texture measurement techniques. The descriptions are divided 
into ti>70 main categories: contact methods and noncontact methods. The 
techniques described below produce measures of a wide variety of pavement 
texture characteristics including: average texture depth, mean void width, 
root mean square of the texture depth, the root mean square of the texture 
slope, and outflow time. The results of attempts to evaluate and validate 
methods of ro.easuring pavement macrotexture are reported below. Many in- 
vestigators have attempted to validate macrotexture measurement methods 
by conparisori with pavement friction coefficients. Most such attempts 
have had very low correlation coefficients. These disappointing results 
are not surprising, because pavement friction is known to be a function 
of bo til microtexture and macrotexture. Therefore, a high correlation 
cannot be expected in an analysis of pavement friction and macrotexture 
data unless a measure of microtexture is also utilized. The contact 
methods of measuring pavement texture are described first followed by the 
noncontact methods« 



12 



1. Contact methods ; The contact methods discussed in this sec- 
tion include: sand patch, sand track, grease patch, putty impression, out- 
flow meter, profile tracing, and tire noise. Altogether, 19 different tech- 
niques are presented under this general heading, 

a. Sand patch methods ; The sand patch method was originally 
developed at the British Road Research Laboratory—' and is one of the first 
methods used to measure pavement surface texture. Since its development, 
several versions of the method have been used by various researchers and 
state highway departments in the United States. Three versions of the 
method are discussed below. These include: (1) the method most commonly 
used by the states and referred to simply as the "sand patchj" (2) the 
modified sand patch; and (3) the vibrating sand patch. : 

(1) Sand patch: The procedure for this method is de- 
scribed in an American Concrete Paving Association technical publication— 
and e 1 sewhere . - ^ i / y -j-j ^ oo / -j-j^g method involves spreading a known volume of 
fine, dry ASTM C-109 Ottawa sand (passing No. 50 and retained on No. 100 
sieve) over a sm^ll area of the pavement surface. The sand is spread with 
a rubber disc into a circular patch until the pavement surface depressions 
are filled to the level of the aggregate tips. The area of this patch is 
determined from an average of diameter measurements taken at four equally 
spaced locations. The average texture depth is calculated as the ratio of 
the volume of sand spread to the area of the patch. The average texture 
depth is the only texture parameter evaluated by this method. 

Mixed results have been obtained from use of the sand 
patch method because of its strong dependence upon possible operator error. 
Hegmon and Mizoguchi^i' reported poor repeatability with sand patch tests 
and noted that good repeatability could be obtained only with extreme care 
by the same operator. 

Ghamberlin and AmslerZ' analyzed sand patch test data 
collected from concrete pavement test surfaces during an HPR study. The mea- 
surements were made in connection with a field investigation of four pavement 
surface texturing methods applied to two sections each of five different pav- 
ing jobs# Sand patch measurements were taken at three different sites within 
each of the 40 test subsections by three different operators performing two 
tests each. Thus, 720 individual sand patch measurements were taken during 
the investigation. 

The analysis of the data produced estimates of the re- 
peatability (method precision) and reproducibility (applied precision) of the 
sand patch test, as well as sampling errors that can be expected in measuring 
the mean texture depths of a pavement section by the method. All three 
measures of the sand patch method were described in terms of the standard 
deviation, ct, of the mean texture depth. Linear regression equations of the 



13 



a = aX + b 



form were derived for each of these quantities for a texture depth, X, range 
of applicability of 0,25 to 2.03 mm (0,01 to 0,08 in.). The coefficients 
determined for each of the three quantities along with the associated correla* 
tion coefficient, r, are given below: 



— b r 

Repeatability in mm (in.) 0.0229 0.0076 (0.0003) 0.30 

Reproducibility in mm (in.) 0.0725 0.0483 (0.0019) 0.42 

Sampling error in mm (in.) 0.1414 0.051 (0.002) 0.57 

Using these results, Ghamberlin and Amsler— also developed nomographs for 
estimating the number of tests an operator would need to perform for a de- 
sired precision and sampling error. 

Some overall findings from the study are worthy of noting. 
Variations in the texture depth within any particular job were found to be 
roughly equivalent to hose between jobs. This suggests a need for positive 
control on surface texturing during constiruction to assure that desired 
texture depths and greater uniformity in texture are attained. Differences 
in repeat tests by the same operator at the same site, account for only 0.4% 
of the total variance. This is substantially less than the 3.37, found for 
the operator-to-operator component of the total variance. 

In spite of conflicting reports about the sand patch 
method, most users of the method consider it reasonably precise, rapid to 
use, low cost and simple. However, the method is not suitable for use on 
grooved and deeply, interconnected textured surfaces, because the sand tends 
to run out along the grooves and channels which makes the true area covered 
extremely difficult to evaluate. 

In July of 1974, a task group in ASTM Subcommittee 
E17.23 was created to develop a simplified method for measuring pavement 
recommended to the subcommittee. The sand patch procedure is currently 
being written up for submission to ASTM as a recommended standard method 
of pavement surface macrotexture determination. 



14 



(2) Modified sand patch : Several modified sand patch 
methods have been used in the past to obtain field estimates of the average 
macrotexture depth of pavement surfaces* ^ ??°^?^^ ^ These methods dif- 
fer from the sand patch method just described in that a volume of sand re- 
quired to cover a specified area is determined, rather than the area that 
will be covered by a predetermined volume of sand. 

In practice, a metal or rubber plate with a cutout of 
knox^ volume is placed on the pavement surface. Fine, dry sand is used to 
fill the cavity. This amount of sand, less the amount required to fill the 
cavity when the plate is on a perfectly flat surface, determines the volume 
of texture below the cavity. The average macrotexture depth is then computed 
as the ratio of the volume of texture to the area covered. 

Dahir and Lentzi^' report that the FHWA has developed a 
modified sand patch method that is suitable for use with pavement core 
samples. This procedure involves installing 6.4 mm (1/4 in.) width rubber 
band around the periphery of the core sample and level with the surface peaks. 
The sample is first weighed and then dry ASTM G-109 Ottawa sand is spread 
on the core surface in a circular motion using a rubber disc. The surface 
depressions are filled to the level of their peaks and the sample is re- 
weighed. The difference in the sample weights represents the amount of sand 
required to fill, the depressions in the surface of the core sanqjle. The 
average macrotexture depth is computed from the difference in the sample 
weights, the specific gravity of the sand, and the circular area of the 
core sample. 

(3) Vibrating sand patch: A vibrating sand patch 

" ' ^ ?'^ Aft/ 

method was developed by researchers at Pennsylvania State University .■^-'? ^°' 
The process was developed to eliminate the operator effect which is one of 
the sources of error in the use of the sand patch technique. In practice, 
a 152 ram (6 in.) diameter pavement core was placed on a pneumatic shaker, and 
sand was added to fill up the space formed by the interior of a ring sealed to 
the exposed texture surface of the core.-'' The weight of sand used was then 
determined. The process was then repeated using a smooth reference surface 
instead of the core sample. The weight of this reference volume of sand 
was subtracted from the first sand weight to determine the weight of sand 
filling the voids in the surface of the core sample. The average macro- 
texture depth is computed from the difference in the sand weights, the 
specific gravity of the sand, and the interior area of the ring sealed to 
the core surface. Repeatability in the texture measurements was greatly 
improved over other sand patch methods and was independent of the operator. 
However, the procedure is suitable only for laboratory use. 

" Hegmon and Mizoguchi-^' report the amount of sand perculating into the 
voids of the core surface was found to be dependent upon frequency, 
amplitude and duration of shaking. The reasons for these dependen- 
cies were not explained. However, these three vibrational parameters 
were controlled during core to core testing. 

15 



^* Sand track ; The sand track device represents another re- 
finement of the sand patch method « -50> 68 / The unit was also developed at the 
Pennsylvania State University and is believed to involve less operator error 
than the sand patch method by controlling the placement of the sand on the 
pavement surface. A constant volume of either ASTM G-190 or G-109 Ottawa 
sand is placed in a trapezoidal hopper and then the hopper is slid along a 
metal frame at a constant rate* Sand flows freely from the bottom of the 
hopper and the unit is calibrated so that it will deposit a uniform thickness 
of 3»2 mm (1/8 in.) of sand on a smooth surface. On a textured surface, 
the unit will deposit sand to fill the surface voids and will place a uni- 
form depth of sand 3.2 mm (1/8 in.) above the peaks of the asperties. 
The resultant test data are reported as sand track values which are lengths 
of travel of the hopper required to deposit the known volume of sand on the 
textured surface. The average macrotexture depth is determined by sub- 
tracting the depth of sand deposited on a smooth surface (3,2 mm (1/8 in.)) 
from the depth of sand deposited on the pavement surface (calculated from 
the kno^vn volume of sand distributed over an area of recorded length and 
known width) • 

The device is adapted to routine field tests and is reported 
to be capable of measuring varying degrees of texture .-^'^ However, the 
precision and accuracy of the sand track device cannot be determined from 
the meager test data reported in the literature. 

c. Grease patch : The grease patch or grease smear method 
was developed by NASA.^^? ^§7 It is very similar to the sand patch method. 
A known volume of grease contained in a tube with a plunger (for ease in 
applying the grease) is placed on the pavement between two parallel strips 
of masking tape. The grease is then worked into the voids using a squeegee 
with a rubber face similar in hardness to an automobile tire. The average 
texture depth is determined by dividing the volume of the grease by the 
area coveredJt§/ The method is quick and easy to apply. Relatively weak 
correlations have been found between average texture depths deteirmined by 
the grease patch method and pavement friction coefficients. No data were 
found in the literature relating grease patch texture measurements and 
other texture measurement methods. 

^* Putty impression : Basically, three putty impression 
methods have been used to make macrotexture measurements. Two of the three 
methods, and the ones most commonly used, resemble the sand patch method in 
application and yield an estimate of the average macrotexture depth.M/ The 
third method involves making a solid casting of the surface for detailed 
laboratory measurements of various texture parameters .i^/ 



16 



i 



The first putty impression method uses a 25 mm (1 in.) thick, 
152 mm (6 in») diameter metal plate and a fixed weight of putty to measure 
pavement macro texture* One side of the metal plate contains a centered recess 
which is 101.6 mm (4 in.) in diameter and 1.6 ram (1/16 in.) deep. The other 
side of the plate is flat and without a recess. To determine the texture 
depth with this method, 15.90 grams of putty (in the shape of a ball) are 
placed on the pavement surface to be measured. The metal plate with the 
recessed side facing the pavement is then centered over the putty and pres- 
sure is applied to flatten the putty between the plate and the surface. On 
a perfectly smooth surface, the recess in the plate will be completely filled 
with the 15.90 grams of putty. As the pavement texture increases, the diam- 
eter of the distributed putty will decrease. The average texture depth is 
determined by dividing the volume of the putty used by the area of the flat- 
tened putty. In practice, the area of the flattened putty is calculated from 
an average of four diameter measurements. The diameter measurements are a 
source of error ^ich directly affect the accuracy of the average macrotexture 
depth calculation. 

A second putty impression method was developed that attempts 
to eliminate this error.^^^/ jn this second method, which is a modified 
version of the first method, a frame, a top plate assembly, a roller, a 0.03- 
mm (0.001 in.) to 0.05 mm (0.002 in.) thick plastic film, a thinner household 
plastic wrap (such as Handiwrap or equivalent) 16 to 20 gramis of silicon putty, 
and a cookie cutter about 67-nHn (2-5/8 in.) in diameter are used to determine 
the average macrotexture depth of the pavement. The frame is a 152-mm (6 in.) 
by 203-ram (8 in.) by 6.4 ram (1/4 in.) thick plate with a 102-ram (4 in.) by 
152-mm (6 in.) center cut of it. The top plate assembly consists of two 
plates joined together. One plate is 152-mm (6 in.) by 203-ram (8 in.) by 
4.8 mm (3/16 in.) thick (the same platform size as the frame) and the other 
is 102-mm (4 in.) by 152-mra (6 in.) by 4.8-ram (3/16 in.) thick and is 
fastened to the top plate so that it fits into the cut out in the frame 
when the top plate completely covers the frame. When the top plate assembly 
and frame are together a space 1.6 mm (1/16 in.) deep exists between a smooth 
surface and the top plate assembly. The roller, which is used to check this 
1.6 ram (1/16 in.) dimension, rides on the 6.4 nm (1/4 in.) thick frame and 
also fits into the frame cut out and penetrates the opening by 4.8 ram (3/16 
in.) leaving the 1.6 mm (1/16 in.) space. 

In practice, the average texture depth is determined by the 
second putty impression method in the following manner. First, the frame 
is placed over a smooth piece of shim stock. Next, putty with a known 
specific gravity is placed on the smooth surface and inside the metal 
frame. The putty is then flattened by covering it with the plastic film 
and pressing it with the top plate assembly until contact is made between 
the plate, frame, and the shim stock (which should sit on top of the sur- 
face asperities and conform to large surface variations). The roller is 
then passed over the flattened putty to check if the putty was pressed down 



17 



properly* The plastic film is then removed and the 67 ram (2-5/8 in.) diam- 
eter cookie cutter is then used to obtain a sample of the flattened putty 
for weighing to the nearest 0.01 grams* The thickness of the putty is then 
determined using the specific gravity, the area and weight of the cut out 
sample of the putty. The same test is repeated again but, instead of using 
the smooth surfaced shim, the putty is placed on a piece of household 
plastic wrap which covers the pavement surface to be tested. The dif- 
ference in indicated thicknesses obtained from using the shim stock and 
household plastic wrap is the average macrotexture depth. 

Measurement errors are reduced using this second method be- 
cause large surface variations are cancelled when the results of the first 
test, which include the variations, are subtracted from the results of the 
second test. Inaccurate estimates of the area of approximate circles of 
putty are also eliminated by using the cookie cutter, with a fixed area, 
to obtain the sauries. ^/ 

The third putty impression method involves making a casting 
of the pavement surface using an RTV silicone which is an elastic solid 
when dry. A mold of the pavement can be cut into thin sections on a meat 
slicer. The slices can then be analyzed in a laboratory to determine such 
quantities as surface void sizes, void distributions, asperity sizes and 
distributions and to obtain typical surface profiles.—' 

Each of the putty impression methods has some disadvantages. 
The first putty impression method is reasonably fast but it is not very 
accurate. The second method effectively eliminates errors due to large 
scale pavement variations and inaccurate diameter measurements; however, 
it does take longer in the field to make the necessary measurements and 
is much more complicated. The third method also takes longer in the field 
than the first method because the RTV silicone has to cure before it can 
be removed. However, the third method has an additional advantage over the 
first two methods in that a more detailed analysis of the inpression speci- 
men can be made under laboratory conditions. 

Although only limited information is available on the cor- 
relation of putty impression results with skid resistance or various texture 
parameters, Ledbetter and Meyer— ' found that the putty impression texture 
depth is related to the sand patch texture depth by the following equation: 

''^'^sand patch ~ 0.8185 TKOp^-j-y impression* 

where TXD = texture depth. This regression equation has a correlation co- 
efficient of 0.98 and was determined using 400 observations from two PCC 
tests sections and 44 observations on laboatory blocks with various fin- 
ishes* The test section had texture depths ranging from 0*371 mm (0*0146 
in.) to 1*892 ram (0*0745 in*)* 



18 



e» Outflow meter ; Two basic types of outflow meters are in 
general use today. Both types employ the use of a hollow cylinder with a 
rubber ring affixed to the bottom face. In application, the cylinder is 
placed vertically on the pavement surface and loaded so the rubber ring 
contacts the surface asperities in a manner similar to that of an auto- 
mobile tire. The cylinder is then filled with water and the time for the 
water to drain between two reference marks on the cylinder is meausred. 
The two types of outflow meters differ basically only in the manner in 
which the water is allowed to drain from the cylinder. The first uses what 
is called the static drainage method wherein water drains from the cylinder 
solely under the influence of gravitational forces. Water in the cylinder 
is under atmospheric pressure. The second type uses a pressurized drain- 
age method. In this method, the air above the water in the cylinder is 
pressurized to a constant amount above atmospheric and is used to force 
the water out between the rubber ring and the test surface. 

The static drainage outflow meter uses a transparent cylin- 
der which is typically 127 ram (5 in.) in diameter and about 305 mm (12 in«) 
tall •ll->2^.i^' The bottom of the cylinder has a rubber ring affixed to it. 
Originally this ring was only 6.35 mm (0.25 in.) wide but was made wider when 
it was found that the rubber ring must be wider than the mean surface texture 
asperity spacing to produce useable outflow times. The rubber ring contacts 
the asperities in a manner similar to a tire when the cylinder is loaded with 
metal ring weights. Various rubber compositions, rubber ring widths, and ring 
weights have been used by researchers .£3/ 

Two pretesting procedures have been used by researchers 
when testing with the static drainage outflow meter. First, the rubber 
ring is conditioned to the pavement surface texture by allowing it to set 
on the surface about 4 min before the tests are conducted.^' Secondly, 
the test surface is flooded with water for approximately 1 min before test- 
ing.i£/ Following these steps, the cylinder is filled with water to a cer- 
tain level and the time for a definite volume of water to drain out of the 
cylinder is measured. The timing can be accomplished manually using a stop- 
watch and to time the water as it passes two marked positions, usually in 
the middle one third of the cylinder. The outflow times can be measured to 
the nearest tenth of a second with this method, but there are inherrent er- 
rors associated with the manual measurements. Errors in judgment in stopping 
and starting the stopwatch can be eliminated by installing electrodes in the 
wall of the cylinder to electrically stop and start precision timers. The 
electrodes are capable of determining the time to the nearest milli-second. 
However, outflow times measured to the nearest tenth of a second are con- 
sidered satisfactory.—' The outflow times of the static drainage method 
usually vary from less than 10 sec for an open-graded texture surface to 
greater than 200 sec for very fine textured surfaces.—' 



19 



A very similar static drainage outflow meter was developed 
by the Transportation and Road Research Laboratory (TRRL) .~' The only 
differences between the use of the TRRL device and the one just described 
are that the pavement surface is not pretreated by flooding before testing 
and rather than allowing the water to flow out of the base during filling, 
a plug is used to hold the water and to release it when actually testing 
the surface. The TRRL also tested an outflow meter which uses an ellipti- 
cal rubber plate similar to the shape and size of a tire footprint. This 
unit is used to detect an anisotropy in the drainage characteristics of a 
surface .ilx' 

The second type of outflow meter uses a pressurized drain- 
age method to force the water through the base of the meter. The develop- 
ment of this device grew out of a number of modifications made to the static 
drainage outflow meter. When the width of the rubber ring used with the 
first type was increased to make it wider than the mean asperity spacing, 
the contact area on the pavement surface was increased from 1,930 to 7,100 
sq mm (3 to 11 sq in«)» To maintain the same contact pressure over the 
larger area the load had to be increased by a factor of 4« The use of simple 
weight rings was now too cumbersome so the outflow meter was pneumatically 
loaded. The loading was later increased to around 159 KPa (23 psi) which is 
typical of the pressures in a passenger car tire» It was felt that this pres- 
sure would be more representative of actual tire-pavement interactions* This 
increased pressure caused the rubber ring to have a tighter seal with the sur- 
face asperities v^ich increased the outflow time beyond practical limits* To 
reduce the outflow time, the top of the cylinder was sealed and pressurized 
and a water input line was added to allow filling of the cylinder prior to 
the test. During testing with later units, the air pressure is maintained 
below 138 KPa (20 psi) to keep the meter from lifting off of the ground. 
Typical outflow times range from 0.7 seconds to 5.0 seconds. 

The outflow meters which have been described can be used to 
estimate a pavement performance parameter termed the mean hydraulic radius 
(MHR). This parameter is a measure of the pavement surface drainage capa- 
bility and is a function of the macrotexture and porosity of the pavement 
in addition to the properties of the outflow meter and fluid used in the 
tests. An empirical equation for the determination of the MHR is given by 
Moore •£2.' 

Several sources of error common to both types of outflow 
meters haye been investigated by various rese archer s.-iLtiliiit' Gustafson — ' 
investigated the effects of water temperature on outflow times. He found 
that the temperature dependence is negligible on coarse pavement textures 
but is quite strong on smooth pavement starfaces for the temperature range 
of about 3''C to 38'C (38'F to 100*F).-^/ 



20 



Other sources of error include the inclination of the out- 
flow meter and various outflow meter-pavement interfacial parameters. Large 
deviations from the vertical in the inclination of outflow meter can produce 
significant errors in the volume of water that is timed. However, for cyl- 
inder and/or pavement inclination of less than 15 degrees from the vertical, the 
change in the water volume is less than 4% and is considered to be an in- 
significant variation .^ii.' The other parameters which affect the device are 
the water pressure, contact pressure, and the dimensions and characteristics 
of the rubber ring. Additional work is needed to explore these variables and 
their effects more fully. Most of the recommendations to date indicate that 

the outflow meter is not suitable for use in the field because it is cumber- 

31 / 
some, temperamental and generally not very practical^^' nor is it directly 

useful in establishing pavement design specif ications.2_t' Also, it is not 
possible to ascertain the mean hydraulic radius from other texture measure- 
ment methods for purposes of conqjarison.^' 

The static drainage type outflow meter has shown variable re- 
peatability ranging from excellent for rough textured surfaces to unaccept- 
ably poor on smooth textured surfaces. ? . ^ The outflow meter has better 
repeatability than the sand patch method and is more sensitive and reliable 
for pavements that are not too smooth. ^3/ ^ graphical relationship between 
estimated average texture depth and outflow times was reported in a study 
conducted by the State of Colorado and is reproduced in Figure 3»^/ The 
outflow times shown are for 620 ml (48.9 cu in*) of water to flow beneath 
a 95.25 mm (3 .75 in«) I«D. rubber ring with an average head of 159 ram (6«25 
in.) of water* The average texture depths are not measured at the same loca- 
tions as the outflow meter measurements but are estimated measurements* 

The mean hydraulic radius can be determined from the pres- 
surized drainage outflow meter results with a precision of 1.57.. No cor- 
relations were found between the pressurized drainage outflow meter and 
other texture measuring devices. 

Currently six state highway departments are participating 
in a study sponsored by the Implementation Division of FHWA involving the 
field evaluation of the static drainage outflow meter. The states involved 
are Colorado, Louisiana, Mississippi, Nebraska, New Mexico, and West Virginia. 
The data being collected under this study will be correlated with pavement 
skid resistance and/or skid resistance-speed gradients under a variety of 
pavement and weather conditions. Based upon the results of the study, the 
states are to recommend to the FHWA improvements in the device and items for 
the preparation of an effective field procedure for the use of the outflow 
meter. 



21 



Oi- 



2.0'- 







0) 






E 




_c 






E 




c 






. 


0.5 







02 


_c 




•« 






■*- 




_c 






a. 










<u 




a. 






Q 










Qi 










u 




(U 






3 




L- 






4- 




3 






X 




.1- 








1.0 







04 


V 










O) 




0) 






o 




C3> 






u 











4) 




w 






> 




(U 






< 




^ 






-T3 










'4— 


1.5 


(U 






£ 




■4- 

D 





06 


u— 




_E 






UJ 




LU 







0.08 




10 100 

Outflow Time, Seconds 



1000 



Source: Reference 8. 



Figure 3 - Graph Showing Approximate Correlation Between Average 
Pavement Texture Depth and Outflow Time 



22 



* 



f. Profile tracing ; Several types of profile tracers have 
been developed and used for making both macrotexture and microtexture mea- 
surements of pavement surfaces. Five types of macrotexture profile tracers 
are described in this section. These are: (1) the prof ileograph (or pro- 
fileometer); (2) modifed versions of the profileograph; (3) a unit developed 
by the University of New South Wales; (4) the linear traverse device; and 
(5) the Texturemeter. The first three macrotexture profile tracers use a 
stylus which scribes the pavement surface and, through the use of linkages 
or transducers, reproduces and magnifies the motion of the stylus on record- 
ing paper. The linear traverse method uses a microscope whose focus is con- 
stantly adjusted as it is moved across the pavement surface. The movement 
of the focus control is used to plot the profile of the surface on record- 
ing paper. The Texturemeter uses a series of rods with a string passed 
through them. These rods are pressed down onto the pavement surface. As 
the rod movements conform to the surface, they cause deflections in the 
string which is indicated on a dial gage calibrated to read average tex- 
ture dep th • 

The five types of profile tracers identified above are all 
employed to measure pavement macrotexture. The profileograph device can 
also be employed, with appropriate modifications, to measure pavement micro- 
texture. Therefore, for the sake of completeness, two types of microtex- 
ture profile tracers are also described in this section. These are: (1) 
the Gould Surf analyzer ; and (2) the Surfindicator . Both of these units use 
a stylus to record the microtexture. 

Each of the profile tracers is described more fully in the 
following subsections. These descriptions are followed by a discussion of 
the distinctions between microtexture and macrotexture profiles, the sources 
of errors in the profile tracers, and finally, the texture measures deter- 
mined from profile tracing data. 

(1) Macrotexture profile tracing 

(a) Profileograph or prof ileome ter : This device 
was developed by the Texas Highway Department and is strictly a mechanical 
device for evaluating the macrotexture of pavement surfaces. It is designed 
to scribe a magnified profile of the pavement surface texture as a motor 
driven feeler probe is drawn across the surface. A mechanical linkage sys- 
tem magnifies the vertical movement of the probe, and the resulting profile 
is drawn on a chart. Upward vertical excursions of the probe are also re- 
corded on a counter as the cumulative vertical peak height over the length 
traversed by the probe .ZliJ^Zi^/ A reading of 29 on the counter is equi- 
alent to 25.4 ram (1 in.) of cxmulative vertical motion of the probe. Therefore, 
cumulative peak height of the asperities, in inches, is obtained by divid- 
ing the counter reading by 29. Average peak height is obtained by dividing 
the cumulative peak height by the number of peaks. A peak is arbitrarily 
defined as any magnified asperity profile with a minimum height of 1.6 mm 
(1/16 in.) and a maximum base length of 6.4 ram (1/4 in.) or any combination 
of these dimensions with the same total magnitude. 

23 



Microscopic damage to surface being traced has been 
noted with this device. The probe tip has been noted to shear off tips of 
weak asperities. These conditions plus the hysteresis losses and sluggish 
motions associated with the mechanical linkage are suspected sources of error 
in the measurement technique. 

(b) Modified versions of the profileograph ; Addi- 
tional profile tracers to measure macrotexture have been developed by 
Pennsylvania Transportation Institute (PTI) and others.-*-^ ^■'•^*^^?. .-^^ These 
units are similar to the profileograph described above in that they still 
use a motorized drive to draw the stylus across the pavement surface. How- 
ever, these versions use a transducer to transform the stylus motions into 
electrical signals which drive the profile plotter. Two types of transducers 
can be used: a displacement transducer that produces signals proportional 

to the texture height or a linear velocity transducer that produces a signal 
proportional to the profile slope. Comparisons of the results from both 
types of transducers show that the integration of the velocity transducer 
output was "nearly indistinguishable" from the displacement transducer out- 
put*iZ' The stylus of these units has a diameter of 1.60 mm (0.063 in.) 
with a cone angle of 20 degrees to 22 degrees and a tip radius of 0.125 mm 
(0.00492 in.). The stylus is oil quench hardened to a Rockwell hardness 
of 60 and has a life of around 100 passes of 2,662 mm (11.81 in.) long at 
a velocity of 3.91 mm/sec (0.154 in. /sec). 

Profileographs have been found to be migged, easy 
to use, rapid in profile tracing and to compare favorably with other pro- 
file tracing devices in resolution and output presentation.— In addition, 
they can produce profile tracings with a high degree of reproducibility. 19/ 
The precision of the PTI unit was found to be 0.5% of full scale. Profileom- 
eter data collected from 41 test surfaces in Texas were found to be poorly 
correlated with skid resi stance .^^^JL^/ The test pavements consisted of 21 
hot-mix asphalt concrete surfaces, 9 port land cement concrete surfaces, 9 
surface treatments, and 2 seal coat surfaces. The macrotexture depth of these 
surfaces ranged from to 1.8 mm (0 to 0.07 in.). 

An extremely high correlation of 0.9463 was found 
between profile data and sand patch data*22i' These results were obtained for 
a limited nximber of dense graded asphalt surfaces (20) that had a macrotex- 
ture range of from 0.10 to 0.84 ram (0.004 to 0,033 in.) as determined by the 
sand patch method. 

(c) University of New South Wales unit ; Another 
type of profile tracer which uses a transducer was developed by the University 
of New South Wales»2lii' It consists of a counterbalanced lever arm with a 
diamond tipped stylus with a radius of 0.01 mm (0.0005 in.) mounted in one 
end. The arm sits on a pivot and as the arm moves up and down a rod, posi- 
tioned immediately above the arm, is also deflected. This rod is attached to 
a moving anode triode electro-mechanical transducer. The output of the 

24 



transducer is monitored for current changes vrfiich have been determined to be 
proportional to the stylus deflection. A chart recorder was used to record 
the stylus deflections on a chart drum. The drum surface had tape with the 
sticky side facing out. Fine thread was fed through the chart recorder pen 
and adhered to the sticky drum as the pen carriage traced the profile of the 
pavement. When a trace of a fixed or known length was completed the thread 
was removed and measured. The ratio of profile length (thread length) to 
center line length (length of trace) provides an estimate of the average peak 
height of the asperities. 

(d) Linear traverse device ; The linear traverse 
device consists of a motorized lathe and a stereo-microscope with the shaft 
of a potentiometer attached to the microscope focusing shaft.ZAi^' The 
potentiometer is fixed to the microscope and is adjusted with the focus. 
The output of the potentiometer is fed through an amplifier to a strip chart 
recorder. In measuring texture, the pavement specimen is moved transversely 
under the microscope while the operator manually keeps the microscope in con- 
stant focus on the surface of the specimen. Focusing on the varying surface 
elevation results in an aii5)lified tracing on a chart of the surface texture of 
the specimen. It is then necessary to make measurements of this tracing to 
produce measures of the pavement macrotexture. Sources of error are asso- 
ciated with the manual focusing and measurements of the tracing. No data 
were found in the literature relating the linear traverse macrotexture mea- 
surements and other macrotexture measurement methods. 

(e) Texturemeter ; The Texturemeter was developed 
by the Texas Transportation Institute to measure the average macrotexture 
depth of a pavement surface in <:;i i-ii- ^5»67,68 / -jhe device consists of a 
number of evenly spaced vertical rods mounted in a frame. All but two of 
the rods can be moved vertically, against spring pressure, independent of one 
another. One rod at each end of the device is fixed for support. Each 
movable rod has a hole through which a taut string is passed. One end of 

the string is fixed to the Texturemeter frame and the other end is attached 
to the stem of a dial gage extentionmeter mounted on the frame. In appli- 
cation, the rods are held in a vertical position with their ends resting 
against the pavement surface. If the surface is smooth, the string will 
form a straight line, and the dial will read zero. Any measurable irregu- 
larities in the surface will cause the string to form a zig-zag line and 
will result in a dial reading from which the average peak height can be 
calculated or directly indicated. The coarser the pavement macrotexture, 
the higher the dial reading. The TTI Texturemeter uses a 0.03 mm (0.001 
in.) dial gauge and has rods spaced at 15.9 mm (5/8 in.) intervals over a 
254 ram (10 in.) length. A similar device, the Rainhart Text-Ur-Meter uses a 
0.03 mm (0.001 in.) dial gauge with 29 probes spaced evenly over 254 mm 
(10 in.).lZ' Both instrvraients are noted to be simple to operate in the field. 



25 



(2) Micro texture profile tracing ; The profile tracers 
discussed above have been used to measure the macro texture of the pavement. 
A macro texture profile tracer using a stylus and transducer to produce a 
trace of the surface can be converted to measure micro texture by changing 
the stylus size, the stylus traverse velocity, and the transducer sensi- 
tivity. Two type of microtexture profile tracing devices that have been 
used are briefly described below. 

(a) Gould Surf analyzer : The Gould Surfanalyzer 
(Model 150) has been used to measure microtexture of pavement surfaces under 
laboratory conditions *±2.' It consists of a stylus probe, some associated 
electronics, and a display for roughness readings. The stylus has a radius 
of 2»54 X 10"^ ram (1 X 10""^ in#) with a 90 degree cone angle and a traverse 
speed of 2.54 mm/sec (0.10 in. /sec). The sensitivity of the device 

is such that a maximum vertical movanent of approximately 1.3 mm (0.05 in.) 
is permitted. No additional data could be found in the literature on this 
device. The device has a precision of 17, of full scale. 

(b) Surfindicator : The Surf indicator is another 
device that has been used to measure very fine-scaled textures of pavement 
surf aces. i^^J'^/? ^"^ It was originally developed to measure surface textures 
of machined surfaces. It consists of a surface datum pickup with a stylus, 
some associated electronics, and a dial gauge for displaying roughness from 
2.54 to 2,540 10^ ram (1 to 1,000 10"^ in.). The stylus has a conical diamond 
tip with a radius of about 0.01 ram (0.0005 in.) and is permitted to move 
vertically a maximum distance of 1.5 ram (0.06 in.). Changes in the speed at 
which the stylus traverses the pavement surface causes errors in the read- 
ings. Consequently J the traverse speed must be controlled. 

(3) Distinctions between microtexture and macrotexture 
profiles : Because most microtexture measuring devices are also capable of 
measuring small macrotexture features, some technique must be employed to 
isolate or eliminate this data.,±^x=Z' One technique, utilized on the Surf- 
indicator, is to install a small foot at or around the stylus which rides 
on top of the large asperity features. The microtexture is measured from 

a reference point that varies as the macrotexture features vary. This is 
extremely important with the Surfindicator its direct readings would include 
the influence macrotexture in the results had it not been eliminated. 

Most of the profile tracers discussed use transducers 
which produce electrical signals that indicate the texture using the magnitude 
and frequency of the output. A filter can modify the transducer output al- 
lowing only microtexture signals to be provided,—' However, this requires 
some decision to be made about the cutoff points for micro texture^or macro- 
texture sizes. In a recently completed study involving a computer evalua- 
tion of pavement texture, it was determined that texture less than 0.50 mm 
(0.020 in.) could be considered microtexture and values greater than 0.50 mm 



26 



(0.020 in.) could be classified as niacrotexture*£LZ' These same findings were 
verified by researchers at the Pennsylvania Transportation Institute when pave- 
ment profile tracers were subjected to power spectral density analysis .JJ:.2±£' 
Thus, based on these findings, a filter which provides a cutoff frequency of 
2,000 cycles per meter (roughly equivalent to one cycle per 0.02 in.) can be 
used to separate microtexture from macrotexture in the profile tracing. For 
instance, a filter passing only frequencies greater than 2,000 cycles per 
meter would eliminate macrotexture influences and yield only a microtexture 
profile tracing. 

(4) Sources of error in profile tracings ; Profile 
tracings made using a stylus can contain errors due to the geometry and 
dimensions of the stylus. For instance, the contact point of che conical 
probe tip will vary if the tip is too large to penetrate to the bottom of 
some of the voids. A prof ileograph using a stylus with a tip radius of 
0,125 mm (0,00492 in,) and a cone angle of 20 degrees will not be able to 
trace curvatures less than 0,125 ram (0,00492 in,), and slopes greater than 45 
degrees will be distorted,i2i2^' Thus, care must be taken to see that the 
geometry and dimensions of the stylus do not fall within the range of texture 
sizes to be measured. One study of profileometer datai^' indicated that 
profileometer measurements errors of 107o or less could be expected due to non- 
linearity of the electronic components, the stylus size and tilt of the stylus 
during the tracing. 

For microtexture measurements, the stylus probe in its 
uncertain movements will indent the surface creating an error of around 1 x 
10-3 nm (0,3937 x 10"^ in,) in the roughness readings. Indentation does not 
present a significant source of error in macrotexture measurements because 
the smallest details measured are generally much larger than the indentation 
depth ,15/ 

(5) Texture measures determined from profile tracings ; 
Some of the profile tracing devices provide a direct measure of pavement 
surface texture through use of a dial gauge or counter. Other devices, such 
as the prof ileographs and the linear traverse produce a recording of the pro- 
file trace which, in turn, must be analyzed to obtain useful estimates of 
the pavement texture. The profile analyses can be accomplished by making 
manual measurements of the profile and performing manual calculations or by 
digitizing the profile for computer analysis. Both macrotexture and micro- 
texture profile tracings can be processed to obtain such variables as the 
root mean square (RMS) of the texture height (depth), the arithmetic mean 
texture depth, and the RMS slope of the texture. Of course, very detailed 
and complex analysis of the profile traces can be performed quickly through 
computer analysis. One study—' used digitizing and computer analysis to 
obtain: (a) values for the root mean square (RMS) of the texture height 
and the first and second derivatives of the profile trace; (b) the autocor- 
relation functions and their first and second derivatives; and (c) the re- 
sults of various types of power spectrial density analyses. 

27 



Profile tracing, with careful statistical analyses and 
data reduction by profile spectral analyses, can be successfully used to pre- 
dict values for the zero intercept skid number (SNq) and the percent skid 
number-speed gradient. "* ? ^~^' Strong correlations have been found between SN-. 
and the micro texture RMS height and between the percent skid number-speed 
gradient and the macrotexture RMS height. 

g. Tire noise ; Studies by Eatoni^' and Mi trey, et al . ,i^' 
have atempted to relate tire noise to pavement surface macrotexture. The 
approach followed in these two studies was to correlate the average texture 
depth of test pavement surfaces, as determined by another contact method, 
with the tire noise in a specific frequency band. The average texture depth 
of a given pavement surface could then be obtained by recording the tire 
noise for these pavements and applying the derived texture correlations to 
the recorded data. 

In order to evaluate road surface macrotexture from tire 
noise, six operating parameters need to be held constant to assure repeat- 
ability in the measurements. They are: 

Vehicle speed, 

Wheel load. 

Inflation pressure. 

Tread pattern, 

Degree of tread wear, and 

Tire size and cons true tion ..- ? ^ ^' 

Two methods have been used to measure tire noise. Both in- 
volve the use of a microphone with a windscreen which is attached to a re- 
cording or measuring device. In the first method, the microphone is placed 
behind the left rear tire, centered 'with respect to the tire tread, point- 
ing directly at the rear of the contact patch area, and set about 203 mm 
(8 in.) from the road surface.—' The microphone output is amplified in a 
variable gain DC amplifier and this signal is recorded on two channels of 
magnetic tape* One channel records the entire microphone output and is 
called the broadband channel. The other channel is used to record a signal 
which has been filtered to allow only the weaker high frequency signals to 
pass. These high frequency signals are separately amplified to provide a 
signal at a high enough energy level to make analysis of these signals easier* 
The filter is used because the low frequency signals are already at high 
energy levels and would saturate the tape if they were amplified. A white 
noise signals of a known sound pressure level (SPL) is recorded periodically 
as a calibration signal. Various types of spectral analyses can be utilized 
to reduce the noise data. 



28 



The second methodi-L' uses a totally different procedure to 
record the tire noise* Rather than continuously monitoring the tire noise 
from the moving vehicle^ the maximum sound pressure level is recorded as 
the car passes a microphone located 4.6 or 7 .6 m (15 or 25 ft) from the pave- 
ment edge and 1.4 m (4.5 ft) above ground level. Readings are taken from 
an SPL meter using an A weighted scale. This method has certain restrictions 
on the test site. The site needs to be free of reflecting surfaces, wind 
speeds must be less than 19.3 km/hr (12 mph), the SPL of the tire noise must 
be at least 10 db above ambient sound levels and the vehicle being measured 
must be the only vehicle in the test area. 

Tests utilizing the first method were run on a test track 
with five asphaltic surfaces and one portland cement surface. Using a pro- 
fileograph to determine the texture depth, it was found that the sound pres- 
sure level increased as the pavement texture decreased. Significant cor- 
relations were found between macrotexture, as measured by the prof ileograph, 
in the frequency band of 25 to 8,000 cpm and tire noise in the frequency 
band of 2,000 to 8,000 Hz. The correlation coefficient, r, for these fre- 
quency bands ranges from 0.70 to almost 1.0. The highest correlations were 
found in the bands of 500 to 1,600 cpm at 2,000 to 4,000 Hz with a range of 
r from 0.90 to almost 1.0. Resolution, the ability to distinquish between 
sites, is excellent at 2,000 to 4,000 Hz. Relative repeatability (precision) 
has been demonstrated and the tire noise data can consistently rank the pave- 
ment surfaces in the same order as ranking them using SN^g and percent skid 
number-speed gradient. The technique, however, does not have absolute re- 
peatability because variations of SPL readings on the same or similar sur- 
faces but at different times and days has been noticed. 

The second method of texture evaluation was conducted on II 
Portland cement surfaces which had five different types of transverse tex- 
turing. Using the sand patch method to determine the texture depths it was 
foxind that the tire noise increased with increasing texture depths (from 0.2 
ram (0.008 in.) to 1.91 mm (0.075 in.)). (It should be noted that this is 
exactly opposite to the findings of the first method on primarily asphaltic 
surfaces.) The precision of this second method was based upon three measure- 
ments (one each at 64.4, 80.5, and 96.6 km/hr (40, 50, and 60 mph)) at four 
test sites and 50 independent measurements at each of the three speeds at a 
fifth location. It was found that, at the 95% confidence level, there were 
significant differences at all three speeds. It is possible that interac- 
tions of vehicle noises could cause difficulty in using this second method. 
For instance, it was found that engine and exhaust noise dominate the vehicle 
generated noises below 56.3 km/hr (35 mph) and that tire pavement noises 
dominate at speeds above 80.5 km/hr (50 mph).^' In light of this, at least 
two of the test speeds may be suspected of providing inaccurate data. 



29 



2. Noncontact methods ; The noncontact methods discussed in this 
section include: stereophoto-interpretation, modifications of the stereo- 
photo-interpretation, light stylus, laser, light depolarization, photoesti- 
mation, and others. The shadow interpretation and white light speckle 
techniques are included in the other category. These two methods have not 
been used to measure pavement surface texture characteristics; however, 
both have potential for application to texture determination. Altogether, 
11 different techniques are presented under this general heading. 

a. Stereophoto-interpretation method ; This method was de- 
veloped by Schonfeld of the Ontario Highway Department to estimate the skid- 
resistance of pavement surfaces by analyzing stereophotographs of the sur- 
face texture." ? ' The method has undergone a number of modifications 
since its conception. Wnat follows is a summary of the latest published 
description of Schonfeld 's method.^' 

In practice, stereophotographs of approximately 101 mm (3 #94 
in.) square sections of pavement surfaces are obtained with a specially de- 
signed camera/box arrangement* The field equipment used is similar to that 
used in the stereophotogx'aphy method developed in Europe for measuring pave- 
ment texture .^l/ xn the Schonfeld method, a 35 mm single reflex camera with 
a focal length of 55-mm is used for taking pairs of stereophotographs. The 
camera is mounted on a light tight box 457 ram (18 in.) above the pavement. 
The box is equipped with an electronic flash unit that illuminates the 
photographed area at an angle of about 45 degrees. The camera is attached 
to a sliding seat for taking pavement photographs from two positions, about 
95 ram (3.75 in.) apart. A horizontal and vertical reference scale, included 
in the field of view, is also photographed. 

The pavement stereophotographs are viewed under a mirro- 
stereoscope or under a raicrostereoscope . If a mirro-stereoscope is used, 
the 35-mm pavement photographs are first enlarged into natural scale black 
and white prints. These are then viewed at a magnification of 6. A micro- 
stereoscope with a magnification of 25, can be used for viewing the 35-ram 
color transparencies. 

Texture elements of the pavement surface are classified 
visually and are rated subjectively according to an established severity 
rating for each of six pavement surface parameters. Four of the parameters 
are used to describe the macrotexture characteristics of the surface: the 
average width, height, angularity, and density of the larger aggregates or 
macroparticles. The other two parameters are used to describe the surface 
microtexture characteristics. One of the two is used to characterize the 
background microtexture of the pavement surface; the other one character- 
izes the microtexture on top of the macroparticles. 



30 



The surface texture of a pavement is classified by a texture 
code number which is conq)osed of the set of the six texture parameter num- 
bers. In the stereo-interpretation of the photographs, an operator places 
a transparent grid with 10-ram squares either over one of the prints or under 
the transparency depending upon which is used in the analysis • Ten randomly 
selected squares are marked and ntimbered on the grid* Each of the numbered 
squares of pavement surface is then examined under the stereoscope and the 
number for each of the six texture parameters is recorded. The results of 
classifying the 10 random 10-ram square areas are then combined to yield the 
texture code number for the complete photographed surface area under investi- 
gation* 

Friction weights are determined from correlation plots for 
each of the six texture parameter numbers and associated severity levels. 
The friction weights are skid number increments derived from tests with the 
Ontario Highway Department's locked wheel skid trailer. The photo-interpreted 
skid number of a pavement surface is determined as the sum of the friction 
weights for the six parameters. The method yields photo-interpreted skid 
numbers for test speeds of 49.9 km/hr (31 mph) and 99.8 km/hr (62 mph). 

Holt and Musgrove^' of the Ontario Ministry of Transporta- 
tion and Communications have developed a skid-resistance photo-interpretation 
manual based upon Schonfeld's latest work. The manual describes the method 
of determining and analyzing the six Schonfeld parameters^ as well as ex- 
plaining field sampling procedures, equipment requirements and san5)le cal- 
culations for friction weights and skid resistance numbers. 

Schonfeld, in his earlier work,— ^' reported a correlation 
coefficient of 0.9 between SN's obtained from actual skid trailer tests and 
those obtained from photo-interpretations. However, no estimation of the 
correlation coefficient was available for the same comparisons using the re- 
vised correlation curves. 

b. Automation of the Schonfeld method ; Automation of the 
Schonfeld method was performed by Howerter and Rudd^T./ to remove the human 
subjectivity associated with the visual stereo-interpretation and also to 
provide a more efficient way of implementing the method. This work was 
pursued by ENSCO, Inc., under an HP&R contract with the Maryland State High- 
way Administration. 

Electronic stereophotogrammetric techniques, previously used 
in the field of aerial mapping, were adapted to obtain digital data describ- 
ing the pavement surface from stereophotographs. Comprehensive computer 
algorithms were developed to process these data and to classify the pave- 
ment surface texture automatically. In a demonstration of the automated 
technique using test surfaces, it was shown that the computer algorithms 
yielded Schonfeld surface parameter values that were in reasonable agree- 
ment with those obtained by the manual method. 

31 



Preliminary designs of a camera system, which could be mounted 
on a highway vehicle and collect the required data, were developed and shown 
to be feasible for a moderate expenditure. However, the automated system is 
not cost effective as an operational tool based on current costs for digiti- 
zation of the stereophotographs and for the associated computer processing of 
tlae data. 

Subsequent to this work, Rudd^' reported that ENSCO developed 
a modified automated approach that requires a much smaller set of data than 
the original automated method to evaluate the surface texture. This modified 
method uses a simplified computer algorithm that can yield good estimates of 
the Schonfeld parameters from single-line profile data. This effort was 
undertaken to avoid the use of cumbersome stereophotographs and the need to 
perform off-line stereophotographic digitization. 

A recently completed phase of the ENSCO HP&R contract involved 
the development of specifications for a practical single-line profile scanning 

instrument for the real-time measurement of surface texture from a moving 

11/ 
vehicle.— The output of the instrument would be used in a modified version 

of the automated Schonfeld method using the simplified algorithms. 

In the latter ENSCO study, existing instrument concepts (dis- 
cussed later in this section under laser noncontact methods) were investi- 
gated and evaluated, but proved to be inadequate for use on a vehicle moving 
at 64.4 km/hr (40 mph). A new instrument concept was devised that showed 
the potential for use in the moving vehicle application. The concept in- 
volves projection of a very short duration, slit of light onto the road sur- 
face and detection of the resultant illuminated profile of surface texture 
with a sensitive vidicon television camera. A more detailed description 
of the concept is presented in the report by Gantor^li' 

Analytical and experimental evaluation of this new concept 
have verified its feasibility. A prototype of this instrument is currently 
being built and will be field tested in the near future. 

c. Light stylus ; The light stylus concept was developed as 
a noncontact method of measuring pavement surface texture from a moving 
vehicle traveling at acceptable traffic speeds.^' Basically, three types 
of light stylus devices have been designed and laboratory tested under sim- 
ulated field conditions. 

The first type of light stylus uses a narrow beam of intense 
light which is projected perpendicular to the pavement surface. As the light 
source moves, the projected light spot on the surface will move up and down 
with the texture of the surface. By using a focusing lens and a ground glass 
viewing screen located at an angle, 9, to the horizontal, an image of the 
moving light spot will be visible on the ground glass screen. In theory. 



32 



I 



then, the actual texture depths of the pavement surface can be obtained from 
direct measurements of the vertical displacement of the light spot (as seen 
on the viewing screen) that have been adjusted to account for the image 
magnification and the angle 9. The nature of the light image is such that 
it cannot be converted at this time to an electrical signal for rapid pro- 
cessing. Instead, manual measurements must be made of the light image to 
obtain an estimation of the texture depths. This device has been abandoned 
in favor of the other two types of light stylus devices. 

The second type of light stylus device is a modification of 
the first type in that it uses a projected narrow slit of light instead of 
a light beam. The projected light trace is photographed for later analysis. 
The photographic analysis consists of enlarging the image, measuring the 
areas between major asperities with a planimeter, summing the areas and then 
dividing by the length of the profile to obtain a measure of the texture 
depth. The projected trace can also be analyzed by direct computer analy- 
sis if the profile image is illuminated on a phototube. The masking of the 
project light slit on the pavement by adjacent asperities is a complicating 
factor in this second type. These confused areas are excluded from the 
analysis. 

The third type of light stylus uses what is called the zero- 
slope detector. This device also uses a narrow beam of intense light pro- 
jected perpendicular to the pavement surface. Two photocells are located 
symmetrically on each side of the light source and in the same plane that 
the profile is to be traced. As the light source moves across the pavement 
the reflected light will vary between the two photocells depending on the 
slope direction (positive or negative) of the pavement asperities. At the 
peaks and valleys of the asperities, the light will be evenly distributed 
to each photocell. These points have zero-slopes. If the photocells are 
used as two legs of a wheatstone bridge, the variations of the output voltage 
will indicate the direction of slope and the points of zero slope. Because 
the valleys can be obscured by adjoining aggregate peaks it is advantageous 
to accurately note only the peaks. Using a diode to block the negative 
voltages, the peaks will be the only zero slope points counted. The mean 
void width can be determined by dividing the length of traverse on the sur- 
face of the light spot by the number of peaks. 

The light stylus devices were developed with the intent of 
recording pavement macrotexture profile data from a vehicle moving at high 
speeds. The first device is presently not suitable for practical use. The 
second and third devices are discussed below. 

The second device can be used at high speeds by using a high 
intensity, short duration strobe as a light source. The strobe will provide 
an essentially still profile on the pavement which can be photographed for 
later data reduction in the laboratory. The effects of vehicle bouncing. 



33 



pitching and rolling are a potential problem for the light stylus techniques. 
The focus of the light changes with the height of the source above the pave- 
ment, A solution to this problem for the second type of device has been to 
mount the unit close to a fifth wheel trailing the instrumented vehicle at 
a predetermined distance. This approach appears to filter out the large ve- 
hicle motions and provides an acceptable control of the distance between the 
light source and the pavement. The second light stylus device has produced, 
under laboratory conditions, macrotexture profiles that compared favorably 
with other profile tracers and at a faster, but still relatively slow rate 
of collection. Currently, the second device is not capable of providing im- 
mediate texture depth readings in the field. The planimetering technique is 
the only form of data reduction at present. 

This approach for filtering out large vehicle motions is ap- 
parently not suitable for the third device (zero-slope detector) because this 
unit is more sensitive to the divergence of the light beam. It has been sug- 
gested that a nondiverging light source, such as a laser beam, be used with 
the third device in place of the standard light source. However, no data 
could be found for a unit using a laser beam. The zero-slope detector can 
provide counts of asperity peaks to within "t 57o as they are encountered at 
high speed. These data can be used automatically in the field to provide 
values for mean void width of the pavement texture. 

While the light stylus devices were developed to provide data 
on pavement texture from a moving vehicle, no reports of field test results 
could be found. Likewise, no correlations between the data from the light 
stylus devices and other texture measurement methods are available in the 
literature. 

d. Laser ; There are two basic types of laser pavement tex- 
ture measuring devices. The first of these devices is the Autech Laser 
Dimension Gauge.— i-i— 2-' This unit uses a low power laser to project a small 
spot of light onto the pavement in the normal direction. Two mirrors are 
located symetrically about the projection axis of the laser beam and re- 
ceive the light reflected from the pavement. The two images from the mir- 
rors are then focused onto an image converter. The distance between the two 
images of the light spot on the image converter will vary with the pavement 
surface texture. The image converter measures the time it takes to sweep 
across the two images and converts this information into texture depth mea- 
surements made from an arbitrary reference location. 

The Autech device samples at a rate of 1,000 hertz and aver- 
ages the data over 0.01, 0,1 or 1 sec time periods. '^ At this sampling 
rate and a typical vehicle speed of 64,4 km/hr (40 mph), the device would only 
be able to measure the pavement surface every 18 ram (0.7 in»)»±ii' The low 
sampling rate, therefore, limits the use of the device to stationary or very 

34 



slow vehicle speed applications* From calibration tests and manufacturer* 
data, the device is capable of measuring pavement texture dimensions of from 
0.03 to 1.59 mm (0.001 to 0,0625 in.). 

The second device was developed independently by both ENSCO 
and the British Transportation and Road Research Laboratory .ii' The ENSCO/ 
TRRL device consists of a laser light source and an array of photodetectors 
which is used as the detection device. The light source and detector are 
located symmetrically about the normal to the pavement surface and at an 
angle of 45 degrees from the normal. As the light spot is moved across a 
textured pavement surface, the reflected image will also move across the ar- 
ray of photodetectors. The difference in position of the reflected image 
from a reference location can be determined and these movements are directly 
related to pavement texture variations. 

The geometric configurations of this device will cause data 
"drop outs" in the presence of "steep texture geometries," On wet surfaces, 
false texture depths will be indicated due to the water level in the voids. 
This device does not sample surface textures fast enough to allow it to be 
used from a moving vehicle at high speeds but it does provide useful texture 
depth data when used in a stationary mode. A TRRL report by D, R, C. Cooper—^' 
contains correlation data of this device with the sand patch method. 

e. Light depolarization : The light depolarization device 
was developed by the Naval Ordinance Laboratory for the Federal Highway Ad- 
ministration and was intended to be used for making texture measurements 
from a moving vehicle, . ?m ' ^/ This system involves the use of a linearly 
polarized laser light source which is directed towards the pavement surface 
at some angle from the normal. If the pavement were a perfect reflector, 
no light would be transmitted into the surface and the beam would be re- 
flected completely at an equal but opposite angle to the normal with no 
change in polarity. However, if the surface were less than an ideal re- 
flector and textured, the light will then be reflected back at other angles 
and become diffused or scattered. The composite polarization of the re- 
flected wave will also cease to be linear and will exhibit an elliptical 
polarization. The more textured the surface is, the larger the degree of 
ellipticity will be. It is this ellipticity, or departure from linearity, 
that is used to measure the surface texture. 

For this device, the laser light source is a helium-neon gas 
laser radiating at a wavelength of 0*6328 microns (2.491 x 10"" in.), but 
any visible or near infrared laser would also work.^28/ Th incident wave 
polarization must be constant and the laser must be of sufficient power (in 
this case 10 milliwatts) to provide a useful signal to noise ratio. 

The receiver consists of a rotating polarizing sheet in front 
of a photodetector. The rotating polarizer modulates the reflected light 



35 



according to the degree of ellipticity. The reflecting light passing 

through the polarizer is sinusoidal and produces a sinusoidal output from 

the photodetector. The measure of ellipticity is the ratio of the minimum 

to maximum values of the sinusoidal voltage from the photodetector out- 
put.28^/ 

The light depolarization method in its present form cannot be 
used to directly measure pavement texture; but, it can be used to rank order 
pavement surfaces according to their degree of texture* The device was 
used in the laboratory to rank order a number of test specimens of pavements. 
An area approximately 51 mm (2 in«) in diameter, or a span of about 10 
void widths, was illuminated by the laser light source. The test results 
were then compared with the rank ordering of the specimens using conven- 
tional texturing measuring techniques, including sand patch, texturemeter 
and outflow meter. The rank difference correlations were at least 0.5 in 
all comparisons. The laboratory measurements with the light depolarization 
device were found to be repeatable to within ± 3% of the initial test values. 

The device has also been field tested from a moving vehicle, 
but these data were not available to the authors of this report. However, 
it is known that the test data were not compared with data from other tex- 
ture measurement devices. 

Besides the pavement surface texture, it was found that pave- 
ment color and "material properties" of the surface also affect the amount 
of light depolarization.^' This effect was not quantifiable and further 
tests were recommended to explore the extent and significance of these 
variations. The system did not appear to be significantly affected by 
simulated vehicle bouncing which would be encountered in real-life high 
speed applications. 

f. Photoestimation (MRI Method) ; The photoestimation tech- 
nique was developed by MRI to determine the skid number- speed gradients of 
pavement surfaces from a moving vehicle. A complete description of this 
technique is presented in Appendix C of Volume I of this report. What fol- 
lows is a brief summary of the photoestimation method. 

In application of this method, a high- intensity short-dura- 
tion flash was used to project a beam of light containing shadow lines onto 
the pavement surface. These lines were then photographed with a 35 ram data 
camera. The shadow lines were produced by a photographic glass slide with 
sharp edged opaque bars placed in the light path and focused onto the pave- 
ment surface. The slide was positioned at an angle so that a portion of 
the shadow lines would remain in focus even though the optical path length 
changed as the vehicle bounced up and down. This beam of light with the 
shadow bars was projected at a low incidence angle to the horizontal to 
project shadows across the peaks and valleys of pavement macro texture. 



36 



The illuminated area of the pavement surface was shielded from 
ambient light sources so that pavement photographs could be taken during the 
day and at vehicle speed up to 64«4 km/hr (40 mph)« The field photographs 
were rated on a scale of 1 to 5 by comparing them to standard photographs 
selected through a series of studies using photographs of pavements with known 
skid number-speed gradients* The ratings of the field photographs were con- 
verted to estimated skid number-speed gradient using regression equations* 

The photoestimation procedure was found to have a correlation 
of 0.81 when the estimated gradients were compared to the actual gradient. 
The method also has a reasonable reliability which is shown by a correlation 
of 0,69 or greater between the first reading and later rereading of some of 
the photographic data. 

g. Other noncontact methods ; Two additional noncontact 
methods are described below. One is the shadow interpretation method and 
the other is the white light speckle method. Neither of these techniques 
have been used to measure pavement surface texture characteristics. How- 
ever, both have potential for application to texture determination. 

(1) Shadow interpretation ; The shadow interpretation 
is a photographic method developed by the Ontario Highway Department to 
provide visual records of the road surface wear due to studded tires.^' 
The method is indicated to also be applicable to macrotexture measurements. 

In application, elevation reference studs are counter- 
sunk into the pavement and are used to provide a fixed height support for 
a piano wire strung on a bow shaped angle iron frame. A light-tight box is 
then placed over the piano wire. This box contains a 35-mra camera which 
looks vertically downward at the pavement and a photoflash unit which is 
offset to one side of the wire. The flash unit projects a shadow of the 
piano wire at an angle to show an oblique cross section of the pavement sur- 
face. The shadow is then photographed for later analysis. 

The variations in the shadow of the wire in the photo- 
graphs reflect changes in the pavement texture depth. Estimates of the 
average texture depth can be calculated using profile analysis techniques 
and shadow measurement adjusted for the projection angle. 

No correlation between data from this method and other 
texture measurement methods are available in the literature. 

(2) White light speckle ; The white light speckle 

method was developed to measure the fine-scaled surface texture on machined 

80/ 
metal surfaces without damaging the surface by traversing it with a probe. — ' 

A white light source is made spatially coherent using a lens and the coherent 

light source is used to illuminate the surface being measured. When the light 



37 



is reflected, the texture of the surface will cause the light to have dif- 
ferent pathlengths and to be reflected at some angle depending on the tex- 
ture depth and the texture slope. When the light is collected and refocused, 
the light received at each point in the image will be coming from several 
different points on the object. If these variations are significant, the 
resulting interference patterns will be seen as speckles. 

To provide a useable output for actual determination of 
the texture parameters, a mask with a pin hole in the center is placed be- 
tween the refocused light and a photodetector . The pinhole must be smaller 
than the smallest speckle. It has been determined that the speckle contrast 
increases as texture depth increases which means that variations in the con- 
trast read by the photodetector will correspond to variations in the texture 
depth. A narrow trace is made across a two-dimensional speckle pattern and 
the resulting output of the photodetector can be used to create a profile 
of the surface. By adjusting the pinhole ahead of the image plane it is 
possible to create variations in the width cutoff point to effectively fil- 
ter out unwanted data. This technique shows much promise as a texture mea- 
surement device because "it can separate out the average surface height de- 
viations from other surface characteristics and is flexible enough to allow 

on / 

for some choice of roughness width cutoff. ".2^' 



B. Current State Practice for Macrotexture Measurement 

Sixteen state highway departments were contacted by telephone and 
letters were sent to all the FHWA Regional Offices to determine the methods 
that have been and are being used by the states to measure macrotexture. 
This included techniques used both in the laboratory and in field work. We 
also wanted to know which of the many techniques does the state prefer and 
on what basis. The states contacted by phone were: California, Connecticut, 
Kansas, Louisiana, Maine, Maryland, Massachusetts, Michigan, Mississippi, 
North Carolina, Ohio, Pennsylvania, Rhose Island, South Carolina, Texas, and 
West Virginia. Personal visits were made to six of these states (California, 
Kansas, Maryland, Ohio, Pennsylvania, and Texas) to obtain additional infor- 
mation. Also, follow-up telephone contacts were made with FHWA' s Region 3, 
4, and 7 offices and a personal visit was made with, the highway engineers 
at the Region 7 office. 

A majority of the states surveyed had some experience with tex- 
ture measurement techniques. Those that did have some experience, generally 
used the methods as an experimental or research tool and not as a source of 
data for pavement construction acceptance, surface specifications, or surface 
condition evaluation. The two exceptions to this were the states of Texas 
and Florida. Texas uses the sand patch method in their specifications for 
surface texture requirements of portland cement concrete. Florida uses the 
Text-Ur-Meter results as a variable in their Present Serviceability Index to 

38 



predict the serviceability rating for pavements and in most of their 
pavement condition surveys. 

A brief summary is given below of some of the state's experience 
with texture measurement techniques. General comments by some of the state 
highway engineers on certain measurement techniques are also presented. 

1. California : California has tried a number of texture mea- 
surement techniques. These include: sand patch, modified sand patch, 
stereophotographic, prof ileometer , outflow meter, putty impression, clay 
casting, and resin casting. A study of the repeatability of the sand patch 
method has shown the technique to be operator dependent. They have used a 
stereophotographic method, different than the Schonfeld method, which in- 
volves interpreting a stereophoto of the pavement surface using photogram- 
metric techniques common in the field of aerial mapping. The stereophoto- 
graph covers only 3,900 sq ram (6 sq in.) of the pavement surface and it cost 
about $200 per frame to interpret* 

2. Connecticut : Connecticut has used both the putty impression 
and sand patch methods. Of the two, they feel the putty impression method 
is more accurate where there are extremes in texture and is simpler to use. 
They also like the repeatability of the putty impression results. Connecticut 
has noted three problems with the sand patch method: 

• It does not work well on very smooth pavement surfaces; 

. It does not work well on open-graded pavement surfaces; and 

. The wind distrubs the sand during the testing. 

Connecticut has found that the sand patch method tends to indi- 
cate a deeper texture depth than the putty impression method on extremely 
coarse textures. On open-graded friction courses, the sand tends to perco- 
late down below the reach of the putty. In this situation, they feel the 
putty impression method provides a more realistic measure of the effective 
texture of the pavement surface in contact with the vehicle tires. 

3. Louisiana : Louisiana has used the sand patch, Schonfeld 
stereophotographic, and outflow meter techniques in their research studies. 
Of the three, they consider the sand patch to be the best method for their 
needs. They are one of the six states currently participating in the FHWA- 
Implementation Division sponsored study involving the field evaluation of 
the static drainage outflow meter. Louisiana made the first set of out- 
flow meter measurements under this study in the fall of 1977 with the FHWA 
supplied unit. The measurements will be repeated in April 1978 after which 
an evaluation report will be written. 



39 



4. Maryland : Maryland has been actively supporting, for the past 
several years, attempts to automate the Schonfeld method for highway surface 
texture classification. The work has been done by ENSCO under an HP&R con- 
tract with the state. Recently, the state's interest has been directed to- 
wards the development of speicifications for, and a prototype of a single- 
line profile scanning instrument for measuring pavement surface macrotexture 
from a moving vehicle. This work is also being conducted by ENSCO under con- 
tract with the state and was described earlier in this section of the report 
under automation of the Schonfeld method. 

Maryland does not like contact methods for measuring pavement sur- 
face macrotexture. They feel the methods are too dangerous from a traffic 
control standpoint. For this reason, their interest have been directed to- 
wards methods of making pavement macrotexture measurements from a vehicle 
moving at highway speeds of 64.4 km/hr (40 mph). 

5» Mississippi : Mississippi has used the sand patch, putty im- 
pression, and outflow meter methods as research tools and have no preference. 
In the past, they have not conducted any correlation or repeatability stud- 
ies of the techniques. However, Mississippi is currently participating in 
the FHWA sponsored outflow meter evaluation study. 

6. North Carolina ; North Carolina has used the sand patch method 
only on an experimental basis and has not performed any study of the repeat- 
ability of the technique. 

7. Ohio: Ohio has also used the sand patch method on an experi- 
mental or trial basis and has not evaluated the repeatability of the tech- 
nique . 

8. Pennsylvania ; Pennsylvania has used both the sand patch and 
sand track methods of measuring pavement surface macrotexture. These methods 
have been used as a research tool and are not used as part of pavement sur- 
face texture specifications. 

9. Rhode Island ; Rhode Island has frequently used the sand patch 
method recommended by the American Concrete Paving Association in their . 
Technical Bulletin 19, "Guidelines for Texturing PCC Highway Pavements."— 
The state has developed a statistical sampling procedure for selecting the 
roadway locations for making the sand patch measurements. 

10. South Carolina : South Carolina has extensively used the sand 
patch method and has used the results to compare their typical mix designs. 
However, they do not consider sand patch results appropriate for use in pave- 
ment surface specifications. South Carolina believes the technique to be 
reliable, but has not formally evaluated the method. They use the ACPA's 
recommended sand patch method with one exception — they substitute local sand 



40 



v_ 



for the specified Ottawa sand because of cost considerations. Sand patch 
measurements in the field are conducted in the wheel path of the traffic 
lanes. They have not experienced any particular traffic control problems 
when making the field measurements and they estimate it takes at most 10 
min to make a single sand patch measurement at a given location. 

South Carolina has studied other methods of making texture mea- 
surements, including the sand track method, but they have not used any of 
them because of satisfaction with the sand patch technique. Although they 
favor the sand patch method, they feel it has two disadvantages: 

• The pavement must be completely dry before the texture mea- 
surements can be made. The method will not work if the pave- 
ment is wet. 

• Tests cannot be performed in the wind without the use of a 
wind screen because the sand is susceptible to being blown 
away by the wind. 

l-^* Texas t Texas has used both the sand patch and modified sand 
patch methods. The sand patch test used is described in the Texas standard 
procedure document as procedure Tex 436-A and involves deploying a known 
volume of sand. (The modified sand patch method involves spreading sand 
over a known surface area.) Texas has also tried profile tracing, putty 
impression, the Text-Ur-Meter, and the Schonfeld technique. A report by 
Ledbetter, et al.,itft' presents a linear correlation between results ob- 
tained with the sand patch method (Tex 436-A) and the putty iiqjression 
method (r = 0.98). Texas uses the sand patch method in their specifica- 
tions for surface texture requirements of portland cement concrete. 

12. West Virginia ; West Virginia has used both the sand patch 
and the stereophoto-interpretation methods, but only on an experimental 
basis. The state has not conducted any study of the intercorrelation of 
different texture measurement techniques. West Virginia is also partici- 
pating in the FHWA sponsored outflow meter evaluation study. 

^■^* Kansas : Kansas has used the linear traverse, Schonfeld 
stereophoto-interpretation, sand patch and outflow meter techniques. Both 
the linear traverse and stereophoto-interpretation methods have been used 
in a laboratory study of wear and polish susceptibility of pavement speci- 
mens. The sand patch was used only once on an experimental basis. Kansas 
has used the outflow meter in the field. They do not like to use the out- 
flow meter on dense-graded asphalt pavement surfaces because the outflow 
times tend to be excessive--in some cases up to 15 min. 

14. FHWA Region 3 response ; Pennsylvania, Maryland and West 
Virginia are the states in this region that have the most macrotexture mea- 
surement experience. The activities of each of these three states have been 
discussed. 

41 



15. FHWA Region 4 response : The methods most frequently used in 
this region to measure pavement texture are the sand patch, putty impression, 
grease patch, and the Text-Ur-Meter. The sand patch is used more commonly 
than the other methods and, generally, only as a research tool. For example, 
South Carolina uses the sand patch method results to compare the texture of 
different types of asphalt mixes, but not for pavement construction accept- 
ance, surface specifications, or condition evaluation. An exception to this 
is Florida which uses the Text-Ur-Meter. Florida uses the Text-Ur-Meter data 
as a variable in the Present Serviceability Index to predict the service- 
ability rating for pavements. These data are also used in most all of Florida's 
pavement condition surveys. 

16. FHWA Region 7 response ; The two states most active in this 
region in making pavement surface macrotexture measurements are Kansas and 
Nebraska. The work of Kansas has already been discussed. Nebraska is 
another one of the participating states in the FHWA evaluation study of the 
outflow meter. Nebraska's outflow meter study began in July of 1977. In 
addition to evaluating the outflow meter, they also intend to develop a cor- 
relation between the sand patch and outflow meter results. 



42 



IV. METHODS OF ACHIEVING MACROTEXTURE IN NEW PAVEMENTS 



There are two types of pavement surfaces that can be used to 
achieve a high level of macro texture in new pavements: open-graded asphalt 
surface courses for bituminous pavements and textured surfaces for portland 
cement concrete pavements. Both of these techniques are described in this 
section. Open-graded asphalt friction courses can also be used to overlay 
existing pavement and, therefore, are also discussed briefly in Section V. 
Texturing of new portland cement concrete surfaces is accomplished when the 
concrete is in the plastic state andj, therefore, is discussed only in this 
section. Retexturing of existing portland cement concrete surface can be 
accomplished through pavement grooving and cold milling and a discussion of 
these techniques is given in Section V. 



A. Open-Graded Surfaces for New Bituminous Pavements 

The highest levels of macrotexture depth for new bituminous pave- 
ments have been achieved with open-graded asphalt surface courses. The macro- 
texture depth of an open-graded surface, as determined by the sand patch 
method, is usually in the range of 1 to 3 mm (0,04 to 0,12 in.) as compared 
to less than 0.3 to 3 mm (O.OI to 0«04 in») for conventional dense-graded 
bituminous surfaces* 

An open-graded asphalt mixture is one containing relatively little 
fine aggregate and mineral filler. The standard FHWA specification for open- 
graded asphalt requires between 5 and 15% of aggregate by weight passing the 
No. 8 sieve and between 2 and 5% passing the No. 200 sieve. This gradation 
requirement eliminates many of the fines that would fill in the voids in a 
dense-graded mixture. The increased size of voids provides an open texture 
that allows water to escape more readily from the tire-pavement interface 
providing a smaller skid number-speed gradient* Thus, an open-graded 
surface provides increased skid number at high speeds due to the flatter 
slope of the skid nxmiber-speed relationship* 

Open-graded surfaces have been used for bituminous pavements since 
the late 1940 's in the Western United States. A landmark stvdy, published in 
1968, found that open-graded surfaces had a higher level of skid resistance 
than any other type of bituminous pavement studied.xl' As a result of this 
finding and a subsequent demonstration program sponsored by the Federal High- 
way Administration, open-graded surfaces have been constructed in 49 of the 
50 states plus the District of Columbia and Puerto Rico.^' 



43 



Open-graded surfaces have been known variously as "plant mix seal 
coats," "open-grades asphalt friction courses," "porous friction courses," 
and "popcorn mixes" in the states where they have been used# The multiplicity 
of terms has developed because of the variety of objectives for which open- 
graded surfaces have been used. Originally, open-graded surfaces were used 
as an overlay to seal an existing pavanent surface, hence, the name "plant 
mix seal coat." When the superior friction properties of the material be- 
came apparent, the Federal Highway Administration adopted the name "open- 
graded asphalt friction course." This term is used solely to identify mate- 
rials that meet the FHWA specification discussed below. The Federal Aviation 
Administration prefers the term "porous friction course" for open-graded 
asphalt applied to airport runways. The term "popcorn mix," describing the 
general appearance of an open-graded surface, has been used by some highway 
engineers in a colloquial sense and is not considered a standard phrase. 

1. specification ; Each of the states that have used open-graded 
asphalt surfaces have developed specifications for designing, mixing, and 
placing them. In some states, the specifications have been adopted as a 
published standard. In other states, the specification is still experimental 
and may be varied from job- to- job. 

The macro texture of an open-graded asphalt mixture is determined 
by the aggregate gradation. All states contacted by the authors have chosen 
to control macrotexture by specifying the gradation of the aggregate used in 
the open-graded asphalt mixture rather than by requiring a minimum level of 
macrotexture depth in the surface course. The reluctance of states to spec- 
ify macrotexture depth as an acceptance criterion is based on lack of agree- 
ment on the best method of measuring macrotexture and experience in obtain- 
ing predictable skid resistance and texture from their specified aggregate 
gradation. 

The following master ranges of aggregate gradation for open-graded 
asphalt mixtures are reconmended by the Federal Highway Administration. 



Sieve Size^^ 


Percent Passing^/ 


12.7 ram (1/2 in.) 


100 


9.5 mm (3/8 in.) 


95-100 


No. 4 


30-50 


No. 8 


5-15 


No. 200 


2-5 



al U.S. Standard Sieve Series. 

b/ By volume (this is the same as by weight 
unless the specific gravities of ag- 
gregate being combined are different). 

The recommended aggregate gradation has a nominal maximxmi size of 9.5 mm 
(3/8 in.), but experience has sho^m that up to 5% of material between 

44 



95 mm (3/8 in») and 12«7 ram (1/2 in.) does not change the skid resistant 
properties of the surface. A specific aggregate gradation within the 
recommended master ranges for each sieve size is used for each job. 

A recent survey of state highway departments conducted by the Na- 
tional Cooperative Highway Research Program (NGHRP) found a wide variation 
in master ranges specified by individual states. 2Z' The authors of the NCHRP 
report suggest that these variations result from the need for each state to 
adapt its specification to a locally available aggregate. However, the NCHRP 
survey found the typical gradations used by the states on specific jobs were 
more similar than the master ranges might suggest. The average gradation for 
all of the states surveyed fell within the FHWA specification band. 

Open-graded mixtures require good, durable aggregate to avoid rapid 
loss of their skid resistant properties due to traffic polishing. There are 
two possible approaches to the problem of eliminating soft, polish-suscep- 
tible aggregates from use in open-graded surfaces. The first approach uses 
physical testing to establish durability and polish-susceptibility. FHWA 
recommends that aggregate in open-graded mixtures not exceed 407c, abrasion 
loss using the standard abrasion test procedure (AASHTO T 96). Another phy- 
sical method used by some states is the acceptance of aggregates after simu- 
lated traffic exposure on a polishing machine and satisfactory test results 
with the British Portable Tester. The second approach is to exclude aggre- 
gates on the basis of petrographic analyses of their components. Many, but 
by no means all, carbonate aggregates (such as limestone) are too polish- 
susceptible for use in open-graded surfaces. However, most states have not 
had good success in trying to predict pavement performance from the results 
of petrographic analyses. One procedure that has been used involves the de- 
termination of the percentage of acid insoluble residue which is an indicator 
of carbonate content and, by inference, polish- susceptibility. An interest- 
ing example of the use of the insoluble residue test in specifying aggregate 
for skid-resistant bituminous surfaces in New York is provided by Kearney, 
McAlpin and Burnett.—' Carbonate aggregates that contain less than 10% 
sand-size impurities or mixtures of carbonate and noncarbonate aggregates 
that contain less than 207o noncarbonates are upgraded by blending with 207. 
noncarbonate stone. 

A formal mix design procedure is used to select the asphalt con- 
tent of an open-graded mixture. The design procedure recommended by FHWA 
is described in Appendix A« In this procedure, the asphalt content is de- 
termined by use of the modified California oil equivalent test to measure the 
surface capacity for the course aggregate fraction (passing the No. 4 sieve). 
The surface capacity, K^,, is determined from the percentage of SAE No, 10 oil 
retained when an aggregate sample is immersed in oil and drained. The oil 
equivalent test should be performed with a sample of the same aggregate that 
is planned for use on the job. The appropriate percent asphalt to be used 
in the mixture is determined from the relationship: 



45 



Percent Asphalt = 2.0 K- + 4.0 



In the FHWA procedure, this asphalt content is used in the job mix without 
further testing. 

A somewhat more laborious procedure has been developed by the 
Colorado Department of Highways.^' An estimated optimum asphalt content 
is estimated using the relationship: 



Estimated Optiimam Asphalt Content = 1.5 K + 3.5 



Then, the actual asphalt content to be used on the job is selected from 
laboratory testing of trial mixes below, at, and above the estimated opti- 
mum asphalt content. 

2« Costs ; Costs of open-graded surface courses vary substanti- 
ally throughout the United States. Typical costs for open-graded asphalt 
mixtures were recently provided by the states of Maine, Maryland, Michigan, 
Pennsylvania, and South Carolina. These states provide costs on a common 
basis — cost per ton in place. The average cost per ton for open-graded as- 
phalt mixtures for these five states is $19«90« For example, the yield for 
a 25 .4 ram (1 in.) open-graded asphalt surface course might be 51.5 km/sq m 
(95 Ib/sq yard) (based on 6.5% asphalt by weight, 15% air voids and specific 
gravities of 2«65 and 1.00 for the aggregate and asphalt cement, respectively). 
Based on this yield and the average cost per ton, a typical cost per square 
yard for a 25.4 mm (1 in.) open-graded asphalt surface course in these five 
states is $0.95. The State of California reports placing 25.4 mm (1 in.) 
open-graded asphalt surface courses for approximately $1.44 per sq m ($1.20 
per sq yard). Combining this cost estimate with those of the other five 
states yields an average cost per square meter for open-graded surface courses 
of $1.19 ($1.00 per sq yard). 

The limitation on the polish-susceptibility of aggregates is often 
cited as an economic restriction on the widespread use of open-graded as- 
phalt friction courses in some areas. A number of states do not have suit- 
able aggregate available locally and must import aggregates substantial dis- 
tances to construct open-graded surfaces. This added expense, naturally, 
makes these surfaces less attractive on a budgetary and cost-effectiveness 
basis. However, some states report that substantial aggregate haul distances 
result in only modest increases to the cost of open-graded mixes. For ex- 
ample, one state reports that open-graded surfaces have been placed for 
$1.44 per sq m ($1.20 per sq yard), despite an aggregate haul distance of 
over 804.7 km (500 miles). 



46 



3. Service life : There are no reliable data to compare the ser- 
vice lives of open-graded and conventional dense-graded surfaces. Some open- 
graded surfaces have been in service as long as 12 years with no apparent 
problems. However, other open-graded surfaces have required early replace- 
ment. Many states have only begun placing open-graded surfaces within the 
last 3 years, so these states do not yet have any basis for service life 
comparisons. 

4. Performance ; Most states have found open-graded surfaces to 
perform adequately. In some cases, performance problems such as flushing or 
raveling occurred on the first open-graded surfaces placed in an area. 
Flushing results when the asphalt content of the mix is too high and ravel- 
ing results when the asphalt content is too low or when the aggregate is 
not completely coated with asphalt. However, such problems have largely 
disappeared as pavement designers and contractors have become familiar with 
the material, and the appropriate construction practices. In some states, 
field engineers and inspectors were reluctant to use open-graded mixtures 
since they often require unusually high asphalt contents and relatively low 
mixing temperatures. These concerns have largely disappeared when the pave- 
ment surfaces have performed adequately. The California Department of Trans- 
portation has found that open-graded asphalt surface perform poorly on steep 
grades in severe winter climates where use of tire chains is common. There- 
fore, California has restricted use of open-graded asphalt surfaces to ele- 
vations less than 0»9 km (3,000 ft) above sea level. An excellent discussion 
of construction and performance problems with open-graded surface courses can 
be found in a recent report from Texas by Callaway and Epps»23^' 

An important consideration in effectively constructing an open- 
graded surface is that the open-graded asphalt mixture must be placed over 
both the roadway and shoulders to obtain the full drainage advantages of the 
mixture. If the open-graded surface covers the roadway, but not the shoulders, 
then the flow of water off the pavement is restricted and much of the benefit 
of the open-graded surface may be lost. There is also a consensus that open- 
graded asphalt surfaces must not be placed under ambient temperatures less 
than 15 ,6° G (60'=>F). 

Texas has found that open-graded asphalt surfaces require routine 
preventive maintenance. The asphalt cement binder tends to become brittle 
and needs to be renewed every 9 to 12 months by spraying an asphalt emul- 
sion, known as a "fog seal," onto the pavement. This seal creates a very 
slick surface for 1 or 2 days, but prolongs the life of the open-graded sur- 
face. 



47 



Many engineers were initially concerned that open-graded surfaces 
would produce higher traffic noise levels than dense-graded surfaces. How- 
ever, the experience of many agencies appears to be just the opposite. Kay 
and Stevens—' evaluated the noise generated by three types of tires on pave- 
ments in Arizona, California and Nevada and found that open-graded asphalt 
surfaces produce slightly lower noise levels than dense-graded surfaces, 
Portland cement concrete surfaces or chip-seal coats. 

Open-graded asphalt surfaces were found by Millsdd' to have higher 
skid numbers at 64.4 km/hr (40 mph) than conventional dense-graded surfaces 
in four western states. Subsequent studies such as the work by Adam and 
Shah,—' have also shown that open-graded asphalt surfaces have flatter skid 
number-speed gradients than dense-graded surfaces. These comparative find- 
ings concerning the skid numbers and skid number-speed gradients of dense- 
and open-graded mixes are also illustrated by the test results in Table 1, 
which were obtained by the Texas Department of State Highways and Public 
Transportation.—' Therefore, many proponents of open-graded surfaces see 
them as the best available method to achieve good skid resistance at speeds 
higher than 64.4 km/hr (40 mph). However, some critics contend that the 
differences in skid resistance between open-graded and dense-graded surfaces 
arise primarily because higher-quality (higher microtexture) aggregate is 
often used in open-graded mixes. It is argued, quite reasonably, that good 
skid resistance can also be achieved by the use of high-quality aggregates 
in dense-graded surfaces. The Penn State model for skid resistance, pre- 
sented in Section II, shows that both microtexture and macrotexture have a 
role in providing high skid resistance. Therefore, open-graded surfaces 
should generally have higher skid numbers than dense-graded surfaces made 
from the same aggregate. A rational choice must involve consideration of 
the cost-effectiveness of these alternatives. A formal analysis of the costs 
and benefits of open- and dense-graded surfaces is illustrated in Section VII 
of this report. However, no single analysis can adequately represent con- 
ditions in all states and regions, so determination of the actual cost and 
benefits experienced by individual states should be encouraged. 

5. Advantages and disadvantages : The major advantages and dis- 
advantages of using open-graded asphalt friction courses for new pavements 
are presented in Table 2. Additional advantages and disadvantages that 
apply only to the use of open-graded asphalt friction courses for pavement 
overlays are discussed in Section V. 

The advantages presented for open-graded asphalt friction courses 
show that they perform well in many ways. The disadvantages do not limit 
the performance of open-graded surfaces but do contribute to increased costs. 
For a more complete discussion of the advantages and disadvantages of open- 
graded asphalt friction courses the reader is referred to the recent NCHRP 
Synthesis of Highway Practice, entitled "Open-Graded Friction Courses."^' 



48 






Q 
U 
Q 

O 
I 

m 

CO 

u 

Q 

o 
tn 



CO 

a 
Pi 

H 
CO 
u 

H 









x: 


y*^ 










a 


j: 










^ 


a 

e 


-d- 








\o 




r~ 








• 


o 










\o 


VD 










CJN 


■s^X 




A! 


•-I 










to 


(U 










(U 


> 




,c 


^•^ 




PM 


<u 




Cu 


x; 






^ 


^-s 


X 


a 




(U 




<: 




e 


r-l 


00 


(U 


pq 


u-1 




r~ 


m 


U) 


T3 


• 


o 




^ 


•H 


s-^ 


o 


u-l 




<u 


O 




00 


>«-' 




> 


Z 










< 






J5 
a 


X! 










J«! 


a 










e 


00 








<!• 




vO 








• 


o 










vj- 


■vl- 










vO 


V-^ 





bO MXll I 

« D j3 -H 

^1 4J 4-1 r-l 

(Si X CU '-I 

> <U 0) -H 

< H Q 6 



ml 



vO 



vO 






CO 
1^ 



00 









tOl 










4J 








1 


c 








M 


(U 




(U 


•u 


<u 


•H 




00 


C 


^ 


■o 


rg 


(0 


(U 




<fl 


00 


M 


o 


9 


u 


• 


J* 


u 


z 


o 


o 


> 


(U 








< 


PM 


T3 

•rH 
CO 


-o 

0) 

a 

CO 





1^ 







•u 




*^\ 




1 


c 


X. 


r~- 




u 


(U 


a 


C^l 




(U 


•H 


6 


• 


(U 


^ 


T) 




o 


bO 




to 


z 


^— / 


m 


9 


M 


CO 




u 


z 


O 


^w^ 


r^ 


OJ 








i-i 


> 


•o 


tJ 


x; 


• 


< 


•H 


<u 


Q 


O 




AS 


<U 


^ 






CO 


a. 

CO 

a 


1 

1 


00 




J-l 


vO 




C^l 




0) 


• 


o 






■o 


ON 


vO 






z 


45 


^— \ 






X) 


a 


x; 






•r1 


^ 


a 






^ 




e 


f-l 




CO 


• 


o 


fO 




<u 


o 


m 






at 


00 


%-• 






m 










u 










(U 


x; 


''■^^ 






> 


Ci J3 






< 


• 


o 








n 
•* 






00 







cd 


T) 




-a 






c 


<U 




<u 






o 


Ti 




T) 






•H 


CO 


4J 


CO 






•U 


M 


<0 


t-4 


X 


4-1 


c 


00 


O 


M 


1-1 


tH 


(U 




o 


1 


e 


CO 


> 


0) 




C 


1 


x; 


c 


w 


1-1 


(U 


4J 


D- 


o 


c 


CO 


p. 


o 


(0 


u 


(U 


(U 


o 


SB 


CO 


N--' 


•o 


CO 





4J 
O 

c 
• c 

H CO 
M O 

C u 

o c 

•r< <U 
4J -i-l 

O X) 

<U CO 
CO u 

00 
c 

•H O 

00 
C CO 
cu M 

> o 

•H > 
00 CO 

a» cu 
u x; 

(0 -u 

+J o 
C w 
cu 

•H •> 
T) Cfl 
CO C 
h O 
00 -H 
U 

-o o 

(U cu 
cu C/1 

a 
en 4J 



cu S 
X3 cu 

§> 
CO 

C a. 

T) r-l 
•H CO 
^ M 

Cfl cu 

> 

en 



4J 

C 
CU 

<u u 

a, u-i 

T3 to 

r. <u 

■-I 

4J rO 

c > 

(U 

•H <J-I 
TJ O 

CO 

u u> 

00 (U 
00 
XI CO 
(U M 
CU lU 

o. > 

W cfl 

U CU 
(U S-i 
Xi CO 



•O CO 

•H 4J 

en "1 

•H 

u^ X! 

o u 

tn rt 

§•" 

•H CO 

OJ CU 

•H 3 

c ^ 

•H CO 

U-l > 
cu 

X) <u 
X! 

CU H 
X! 

H 



Cfl 

cu 
00 
CO 
M 
cu 

> 

CO 

0) 
X! 
4J 

e 

o 
u 

M-l 



XI 

o 

x; 
jj 

OJ 

e 

Xi 
CJ 
Vl 
CO 

o, 

XI 

e 

CO 



XI 
CU 

4J >^ 

CO XI 

»-l 

3 T) 

O cu 

r-l C 

CO -H 

^ e 

(U cu 

XI u 

cu 

Q 



col 



^1 



49 



; 



H 

















en 








/-\ 














• 
























<u 


(U 






fe 














QJ 
























o 


M 













73 








O 
























CO 


3 






o 






OJ 








C0« 
























M-l 


4J 






VO 




&, 


T) 








M-l 




















"4-1 




V4 


X 






v»/ 


• 


cu 


CO 




OJ 




5-1 




















O 




3 


OJ 








03 


(U 


5-1 




o 




3 














>> 










0) 


u 






o 


OJ 


^ 


00 




CO 




03 


















c 












o 


-3 




1 




*w 


















1—1 






o 




T3 


4.1 






vO 


CO 


o 


0) 




5-1 




TJ 














•H 






•H 




0) 


3 






• 


U 


4-J 


03 




3 




OJ 














'O 






^ 




^3 


<u 






m 


00 




3 




03 




73 














CO 






CO 




CO 


)-4 






r-4 




>^ 


OJ 








CO 














<u 






U 




S-i 


OJ 








ex 


Vj 


T3 




OJ 




M 














u 






O 
•H 




00 
1 


U-l 






(U 

> 


OJ 
cu 


CO 

03 


M 




x: 

4.J 




00 
1 














■U 




• 


V^ 




<u 


•H 






o 


4J 


03 


O 








3 














o 




>, 


OJ 




OJ 


ts 






X3 


03 


OJ 


<4-l 




5-1 




OJ 














G 




CO 


4J 




OJ 


CO 






CO 


3 


a 

OJ 


3 




0) 

4J 




o 














01 




en 


T3 




T3 








03 


O 


3 


CO 




U-l 


















•I-l 




en 
0) 


-3 




M 


4J 






OJ 
U 


03 


OJ 


x; 

4-1 




CO 




T3 

1—1 














<u 




o 


•H 




o 


3 






3 


3 


-Q 






3 




O 














4J 




<U 


G. 




U-l 


OJ 






u 


•H 




OJ 




0) 


















CO 




C 


CO 






e 






CO 


CO 


t>> 


o 




> 




3 














00 






M 




c 


OJ 






J-" 


jr 


CO 


•H 




cu 




CO 














<u 




en 






CO 


o 






CU 


o 


e 




















OT 




u 




•H 


(U 




^ 


CO 






a 






-a 




4-1 




O 










H 




60 






en 




4.) 


rH 






B 


OJ 


03 


3 




OJ 




4-1 










Z 




00 




<U 


3 






o- 






cu 


M 


4-1 


CO 




3 














Ci] 




fl 


• 


O 


CO 




M 








4J 


•H 


r-l 










>, 










s 






en 


c 


CJ 




0) 


3 








4J 


CO 


;s 




3 




I— 1 










PE] 




JJ 


CD 


CO 






4J 


•H 






4J 




03 


o 




•H 




5-1 










>; 




c 


■U 


c 


en 




U 








3 


>, 




3 




CO 




O 










■< 




CO 


CO 


(U 


iH 




o 


73 






(U 


^ 


00 


03 




E 




o 










P^ 


en 


u 


u 


JJ 


iH 




.fl 


OJ 




• 


•H 




3 






0) 




a 












cu 


en 


<n 


c 


•H 


• 


03 


M 




4-1 


J3 


OJ 


•H 


M-l 




)-i 


• 












^ 


ao 


•H 




•H 


a 


OJ 




•H 




rH 


S 


00 


O 


O 






>i 


OJ 










C4 


CO 


en 


>^ 


CO 


en 


en 


0) 


3 




3 


CO 


CO 


•H 






o 


u 


5^ 










z 


U 


(U 


c 


6 




M 


^ 


3* 




CJ 




e 


OJ 


OJ 




4J 


73 


0) 












c 


i-l 


CO 




4-J 


3 




OJ 




•H 


4-1 


CO 


T3 


OJ 








J3 










p:i 


CtJ 


1 


e 


(U 


o 


O 


>, 


u 




<4-l 


CO 


-3 




5-4 




03 


OJ 


■3 










o 


> 


£ 




> 


3 


u 


CO 






M-l 






OJ 


<4-( 




!-( 


5-1 


CO 










f^ 


-a 


en 


c 


•H 


TS 




S 


03 




•H 


73 


O 


U 






CO 


CO 














ca 


•H 


•H 


4-1 


o 


0) 




•H 




-TD 


QJ 


4J 


O 


OJ 




0) 




03 










CO 


en 


r-i 




C 


S-i 


CJ 


0) 








CJ 




a 


y 




0- 


cn 


>s 










pc] 


•H 


o 


OJ 


Q) 


a 


CO 


>+-l 


^ 




03 


CO 


OJ 




CO 


. 


a 


OJ 


CO 










C/3 


Q 


c:u 


1—1 


> 




14-1 


•H 


)-( 


• 


•H 


1—1 


1— i 


>> 


u-l 


03 


CO 


•H 


^ 










laii 






XI 


OJ 


s 


u 


1—1 


O 


03 




Q. 


XI 


rH 


5-1 


V 




4J 


5-1 










C3 




«^ 


CO 


u 


3 


■ 3 




13 


4J 


00 




•H 


OJ 


3 


o 


4J 


•H 


OJ 










O 




<D 


iH 


a 


tu 


en 


OJ 


TS 


iH 


c 


0) 


4-1 


> 


03 


CO 


3 


5-1 


> 










O 




fH 


•H 




I— 1 




a 


3 


3 


iH 


-Q 


Cu 


•H 




u-l 


OJ 


OJ 


O 














-Q 


CO 


>^ 


o 


<U 


•H 


CO 


03 


,J3 




OJ 


4J 


OJ 


5-1 


E 


a. 












a 




CO 


> 


1—1 


!-i 


J= 


> 


J3 


0) 


O 


4J 


CJ 


CO 




3 


0) 


03 


0) 










o 




M 


CO 


J-i 


4.) 


4-1 


5-1 




!-i 


4J 


03 


03 


r-l 


4J 


03 


> 


CO 


a 










M 




3 




CO 


0) 




OJ 


U-l 




CO 


3 


3 


OJ 






CO 




5 










H 

M 




O 




W 


0-. 




CO 


M 




Pi 


s 


CO 


oi 






Pi 




CO 






















































Pi 




. 




. 


. 




. 


. 




, 


. 


. 


. 






. 




, 










p£4 




r-H 




(>J 


cn 




<■ 


U-l 




vO 


r>. 


CO 


<Ti 






o 

t— I 




•—4 










H 




M 










































hJ 




0) 








1 






1 










1 












1 






^ 




r* 








iJ 






0) 






1 




•H 












CJ 




OJ 




'So 




ti 




OJ 






03 






r- 




TJ 












03 




X 


(li 




•H 




CO 




5 






3 






"cb 


• 


(U 










>. 


3 






W 




-i:= 




J= 






1 




CU 






>H 


OJ 


e 










4J 


OJ 




>. 


< 




en 


, 


jj 




3 
O 


03 




T3 






-^ 


CO 


g 




, 




-3 


•H 

1—1 


-3 




3 
c 


o 




•H 


en 


u 






3 




i^ 






03 


CO 






•3 




OJ 


•H 


!-i 






M 






0) 


(U 




03 


<U 




O 






03 


OJ 


4J 




OJ 


4-1 


-3 


X 


O 




1) 


o 




T3 


o 


5 




QJ 


TT 




M-l 






0) 


Vj 


03 




N 


OJ 


CO 


~l 


u-l 




CI 


3 




0) 


CO 


o 




iH 












i-t 


oc 


O 




•H 


3 


5- 


03 






3 


Od 




(U 


U-l 


rH 


• 


O 


>-i 




3 








00 


E 




£ 




00 


•H 


3 




U-l 


o 




a 


)^ 




en 


•H 







CO 






03 


CO 


iH 




•H 


00 


1 


> 


3 




S-I 


J. 




en 


3 


en 


cu 


^ 


U-4 




x: 






OJ 




CO 




3 


3 


OJ 




'^ 




2 


z 






en 


•H 


o 


0) 






4J 






U 


4-1 




• 


•H 


•H 


03 


4J 


4J 




■J3 


M 




-C 






CO 


> 


3 










•H 


3 


0) 


4J 


E 


5-1 


c 


^ 








(^ 




00 


13 


iH 


4-1 




CO 




>^ 






3 


CO 


o 


3 




3 


aj 


00 


03 




U 


O 


en 


•H 


0) 


CO 


!-i 


S 


r1 




<u 






cr 


iJ 


CO 


OJ 


03 


T) 


73 


•H 


00 








<u 


J= 


T3 


•H 


3 


o 


4-1 




5 






OJ 


03 


U-4 


e 


•H 






3 


3 




Zj 




fcO 




CO 


4J 


en 


u 






o 






!-i 


•H 


5^ 


OJ 




4-1 


^1 




•H 








CO 


■u 


i-i 


c 




4-t 


03 




1-H 








03 


3 


o 


00 


X 


o 


I-l 


v^ 


• 


C 




J-i 


CO 


00 


0) 


TJ 




03 


. 




. 




03 


OJ 


03 


CO 


3 


00 


u-l 


OJ 


5-1 


03 


o 




c 




1 


4J 


OJ 


>. 


<U 


03 


OJ 


03 




03 


^4 




iH 


•H 


•H 




'-; 


3 


OJ 






CO 


0) 


OJ 


o 


X) 


CO 


r-l 


OJ 


M 


<U 




03 


1 


OJ 


a- 


4-1 


3 


3 


4J 


e 


CJ 


3 




> 


o 


en 


D. 


CO 


}-i 




O 


CO 


O- 




3 


T3 


03 




4J 




3 


3 




3 







"^ 


c 


C 




u 


Cu 


03 


CO 




CO 




^ 


•H 


3 


^ 


3 


4-1 


X 


0) 


4-1 


U-l 


•H 




<J 


CO 


0) 


00 


00 


en 


•H 


<4-4 


03 


U-l 


• 


u 


^ 




OJ 


5-1 


CO 


4-1 


3 


3 


S-I 


4-1 






4J 


T3 


fl 


1 






U 


iH 


M 


0) 


•H 


03 


g 


4J 






• 




9! 


3 


3 






en 




•H 


0) 


T3 


03 


3 


0) 


3 


-3 


^ 




CO 


U-4 


x: 


a 


" 03 


u 


e 


03 


S -3 






•H 


)-l 


C 


en 


C 


4.) 


03 


> 


03 


•H 


4J 


>% 


O 


CO 


4-1 


5-1 


5-1 0) 


OJ 


cu 




5-1 0) 






en 


O 


CO 


c 


CO 


3 




0) 




(-1 




4J 






3 


CO 


0) CJ 


3 


> 


T3 


-3 






<u 


M-l 


rH 


0) 




(U 


Xl 


iH 


T) 




e 


•H 


o 


>. 


C- 


T-i 


X 3 




3 


OJ 


U-. Vj 






V4 




a 


TJ 


4= 


6 


CU 




0) 


r-; 


3 


1—1 


•H 


1—1 




00 


4-1 U-l 


5-1 


a 


-3 


3 








c 


o 




0) 


(U 


TD 


OJ 


'O 


4J 


e 


CO 


U-l 


0) 


iH 




3 5-1 


OJ 




3 


4-1 U 






T3 


CO 


^-1 


5-1 


CO 


> 


CO 


03 


CO 


O 


•H 


3 


u-l 


4-1 


o; 


03 


QJ 3 


4J 


U_l 


5-1 


X 11 






•H 


4= 


-3 


O 


r-l 


CO 


Vj 


•H 


S-< 


c 


3 


cr 


CO 


CO 


0) 


03 


3 03 


4-1 


o 


00 


O 5-1 






^ 


4.) 


>% 


iJ-l 


a 


O. 


00 


O 


Of 


e 


•H 




S-i 




r* 


0) 




i) 






'-^ 






CO 




3: 




CO 






z 




CO 


S 




H 




3 


XI 




ca 






(JL, 



ro 



50 



B. Texturing of Portland Cement Concrete Surfaces 

The macrotexture of new portland cement concrete surfaces must be 
established by texturing or finishing the surface of the roadway while the 
concrete is in the plastic state, i.e., before the concrete has hardened. 
Historically, most portland cement concrete pavement surfaces have been 
finished by the burlap drag technique, which is described below. However, 
many other methods for finishing pavements are available. Thirteen alterna- 
tive methods, together with their resulting initial macrotexture depths, as 
determined by the sand patch method are listed in Table 3 to illustrate the 
variety of texturing methods. However, the initial texture depths for some 
finishing methods have been found to decrease rapidly due to traffic wear* 
As state highway departments began perceiving the need for deeper pavement 
texture than could be obtained with the burlap drag, a number of these 
texturing methods were tried. The most commonly used methods have included 
belting, brooming, and wire tining. Most research has identified wire tin- 
ing methods (also called wire combing or plastic grooving) as being the most 
effective in creating deeper texture depths and higher skid numbers. In 
1976, the Federal Highway Administration, recommended transverse wire tining 
as the most practical and dependable method of providing macrotexture for 
new Portland cement concrete pavements. 



TABLE 3 
INITIAL TEXTURE DEPTHS FOR VARIOUS FINISHEsH/ 



Method of Finish Texture Depth (mm (in.)) 

Wood float 0.36 (0.014) 

Light belt 0.38 (0.015) 

Light burlap drag 0.43 (0.017) 

Heavy belt 0.51 (0.020) 

steel wool 0.56 (0.022) 

Heavy burlap drag 0.64 (0.025) 

Wallpaper br^ash 0.66 (0.026) 

Medium paving broom 0.74 (0.029) 

Door mat (cocoa matting) 0.81 (0.032) 

Wire drag 0.91 (0.036) 

Heavy paving broom 0.94 (0.037) 

Flexible wire brush 1.30 (0.051) 

Stiff wire brush 1.91 (0.075) 



51 



1. Specification : This section describes the methods for tex- 
turing Portland cement concrete in the plastic state including burlap drags, 
brooming, belting and wire tining. The other methods listed in Table 3 are 
not described because they are not in general use. 

a. Burlap drag finishes ; Burlap drag finishes are the old- 
est method of texturing PCC pavements. In 1963 about 607. of highway depart- 
ments used the burlap drag method exclusively JLi' In this method, longitudi- 
nal stria tions are formed in the concrete surface by dragging burlap or other 
material behind the paving machine. Refinements of this method have used 
artificial grass or carpet instead of burlap to create a deeper texture. In 
another variation, deeper texture is created by allowing mortar to accumulate 
on the trailing threads of the burlap. 

A typical specification for burlap drag used by the Mississippi 
State Highway Department reads: "A drag shall be used which shall consist of 
a seamless strip of damp burlap or cotton fabric, which shall produce a uni- 
form surface of gritty texture after dragging it longitudinally along the full 
width of pavement. For pavement 4»9 m (16 ft) or more in width, the drag 
shall be mounted on a bridge which travels on the forms. The dimensions of 
the drag shall be such that a strip of burlap or fabric at least 0,9 m (3 ft) 
wide is in contact with the full width of pavement surface while the drag is 
used. The drag shall consist of not less than two layers of burlap with the 
bottom layer approximately 152.4 mm (6 in.) wider than the upper layer. The 
drag shall be maintained in such condition that the resultant surface is of 
uniform appearance and reasonably free from grooves over 1.6 mm (1/16 in.) 
in depth. Drags shall be maintained clean and free from encrusted mortar. 
Drags that cannot be cleaned shall be discarded and new drags substituted." 

' b. Broom finishes : In this method, a broom is used to form 
striations in the pavement surface. Although the burlap drag forms longi- 
tudinal ridges in the plastic concrete, brooming is usually done in a trans- 
verse direction across the pavement. A typical specification for trans- 
verse brooming used by the Connecticut Department of Transportation reads: 
"Brooming operations shall be performed when the water sheen has practically 
disappeared, prior to initial set and before the concrete is in such a con- 
dition that the surface would be torn or unduly roughened. The brooming 
shall extend transversely from edge to edge of the pavement. The broom 
shall be 3 m (10 ft) long and the machine shall be operated in such a manner 
that successive passes of the broom will overlap the previous pass by about 
0.3 m (1 ft)." The types of brooms used include nylon bristle brooms, steel 
brooms, and poly-plastic brooms. 

c. Belt finishes : Like broom finishes, belt finishes are 
used to form transverse striations in the pavement surface. Belts are usu- 
ally made of two-ply canvas and can be used manually to texture pavement 
surfaces. The Mississippi State Highway Department specification for a belt 
finish reads: "...when straight-edging is complete and water sheen has 

52 



practically disappeared and just before the concrete becomes nonplastic, 
the surface shall be belted with a two-ply canvas belt not less than 203 
mm (8 in.) wide and at least 0,9 m (3 ft) longer than the pavement width. 
Hand belts shall have suitable handles to permit controlled uniform manipula- 
tion. The belt shall be operated with short strokes transverse to the road 
center line and with a rapid advance parallel to. the center line." 

d. Wire tine finishes ; Wire tine finishes are placed with 
a steel comb with teeth or tines that are dragged across the pavement sur- 
face. The texture pattern produced in the plastic surface of the portland 
cement concrete is deteirmined by the size and spacing of the tines. Both 
longitudinal and transverse tining have been used by state highway depart- 
ments. A typical specification for transverse tining used in Mississippi 
reads: "The final surface texture shall be produced with a metal tine finish- 
ing device. The tines shall be approximately 0,81 x 2.11 ram (0.032 x 0.083 
in.) steel flat wire, 102 to 127 mm (4 to 5 in.) in length spaced on 12.7 ram 
(1/2 in.) centers. The texturing device shall be so constructed and operated 
as to produce uniform transverse parallel grooves normal to the centerline 
of the pavement and 12.7 mm (1/2 in.) on centers and having a depth of 4.8 
mm (3/16 in.) determined as set out herein below.. .The final depth of the 
finished grooves will be determined by the accurate use of a standard com- 
mercial tire tread depth measuring gauge with 0.8 mm (1/32 in.) graduations 
so arranged as to be easily and accurately read." 

2. Cost ! The cost of texturing of portland cement concrete in 
the plastic state is difficult to estimate because most state highway de- 
partments do not pay for texturing as a separate item. Ledbetter, et al.,i!ft' 
found no cost difference in several experimental finishes that they compared 
and most engineers agree that the difference in cost between texturing methods 
is negligible. Most finishes including longitudinal tining are produced with 
a float machine or curing rig, so construction contractors do not find it 
necessary to purchase a new piece of equipment. Transverse brooming and tin- 
ing miay be an exception to this general mle and may, therefore, involve a 
small additional cost. Since cost is not generally a factor, the texturing 
method that provides the deepest macrotexture, highest skid number and 
lowest wet-pavement accident experience should be adopted. 

3. Service life ; The service life of portland cement concrete 
finishes is difficult to specify exactly because it depends on the depth 

of the initial macrotexture for the type of finish used, the traffic volume 
and the use of tire chains or studded tires. There are, however, several 
sources that provide an indication of the rate at which the texture de- 
grades. For example, the results of a Texas study by Ledbetter, et al.,-t-t' 
summarized in Table 4, indicate the amount of loss in texture depth of 



53 



seven surface finishes over a 30-month period. The study also evaluated 
11 additional surfaces over a 7-month period. The study concluded that 
the texture depth decreased an average of 25 to 35% with traffic wear and 
then leveled off. A Georgia Department of Transportation study reported 
by Thornton^' established a relationship between skid number and cumula- 
tive traffic passages. These results are shown in Figure 4, for six dif- 
ferent surface finishes. Both the Texas and Georgia studies involved pave- 
ment surfaces that are not exposed to tire chains or studs. 



TABLE 4 

44/ 
TEXTURE DEPTH DEGRADATION AFTER 30 MONTHS- ^^ 

Texture Depth (mm (in.)) Percent Loss Over 
of Finish Dec. 7Li ^ June 11^' 30-Month Period 

Transverse broom 1.45 (0.057) 0.76 (0.030) 47 

3.2 mm (1/8 in.) 1.63 (0.064) 1.27 (0.050) 22 

transverse 

tines 
Longitudinal broom 0.91 (0.036) 0.46 (0.018) 50 

3.2 mm (1/8 in.) 2.36 (0.093) 1.30 (0.051) 45 

longitudinal 

tines 
Burlap + 3.2 mm 2.11 (0.083) 1.57 (0.062) 25 

(1/8 in.) 

longitudinal 

tines 
Burlap drag (con- 0.71 (0.028) 0.58 (0.023) 18 

trol) 
Transverse brush 0.79 (0.031) 0.77 (0.026) 16 



a./ Detejrmined by putty impression method and corrected to correspond to 

sand patch method. 
hi Determined by sand patch method. 

4. Perfoinnance ; The Texas and Georgia studies mentioned above 
also provide an indication of the performance of port land cement concrete 
finishes. Ledbetter, et al.,itft' conclude that, under simulated rain condi- 
tions, deep transverse texturing resulted in the greatest improvement in 
skid number. The results of the Thornton study,^' presented in Figure 4, 
show the clear superiority of wire tine finishes over broom and burlap drag 
finishes. Tines provided the highest skid resistance of any finish tested 
and the traffic wear rate for the 12.7 mm (1/2 in.) tine spacing is less than 
any other finish except the burlap with grout build-up, which had much lower 
skid numbers. A report by Balmer— stated that finishes produced by nylon 
brooms, flexible fine wire brushes, wire drags, and modified burlap drags 

54 



o 

-* 

Z 



(U 

E 

z 






6.35mm (1/4") Spaced Grooves (Tine Finish) 
12,7mm (1/2") Spaced Grooves (Tine Finish) 
Heavy Broom Finish 




Cumulative Traffic, Millions 



Figure 4 - Skid Resistance Change From Traffic on PCC Pavements 

55 



met with only limited success* The report also stated that greater success 
was achieved by wire tining, and that it is desirable to finish pavements with 
a burlap drag before tining to produce a gritty texture between the grooves. 

Some concern has been expressed that deeper textures for portland 
cement concrete surfaces would produce higher noise levels* Scarr— ' eval- 
uated nine surface finishes including burlap drag, brooming, and metal tin- 
ing and found no substantial differences in noise levels. Thus, any apparent 
differences in the so\and produced by different surface finishes may result 
from differences in pitch and not from increased noise levels. 

Each study of texturing has concluded that wire tining is the most 
desirable finish for portland cement concrete surfaces. On the basis of this 
strong evidence, the Federal Highway Administration has recommended the use 
of wire tine finishes. Transverse rather than longitudinal tining was recom- 
mended by FHWA because transverse texturing provides higher skid numbers due 
to superior drainage. Recent conversations by the authors with 15 state high- 
way departments determined that all but one of these states have adopted longi- 
tudinal or transverse tining as the standard finish for portland cement con- 
crete. The impact of these new texturing specifications in many states has 
not been great, however, because very few portland cement concrete pave- 
ments are being constructed. One important area where wire tine finishing 
is having an impact is in texturing the surfaces of portland cement concrete 
bridge decks that are replaced. 

5. Advantages and disadvantages: The advantages and disadvantages 
of the four most commonly used methods of texturing portland cement concrete 
bridge decks are presented in Table 5. 



56 



TABLE 5 
TEXTURING OF NEW PORTLAND CEMENT CONCRETE SURFACES 

Advantages Disadvantages 
Texturing Method: Burlap Drag Finish 

1. Requires no additional equipment 1. Provides small macrotexture depth 

2. Provides inconsistent results 

3. Timing is critical 

Texturing Method: Belt Finish 

1. May be done manually 1. Timing is critical 

Texturing Method : Broom Finish 

1. Texture depth superior to burlap 1. May require an additional piece of 
drag finish equipment 

2. Timing is critical 

Texturing Method: Wire Tine Finish 

1. Higher skid resistance then 1. Transverse tining may require an 

other finish methods additional piece of equipment 

2. Lower wear rate than other finish- 2. Tines may dislodge aggregate near 

ing methods the pavement surface 

3. Excellent drainage with transverse 3. Overlaps of tine passes may cause 

tining weak, pavement strips 

4. Timing is critical 



57 



METHODS OF RESTORING MAG ROT EXT URE FOR EXISTING PAVEMENTS 



Existing pavements provide the greatest potential for reducing wet- 
pavement accidents by improving pavement surface macrotexture* Only a limited 
mileage of new pavements are constructed in any given year, but the mileage of 
existing pavements that are candidates for macrotexture improvement is nearly 
unlimited— far greater than the funds available for such improvements could 
treat* This section presents an overview of the four most common methods of 
increasing the macrotexture of pavement surfaces* These are: open-graded 
asphalt overlays, pavement grooving, cold milling, and seal coats. This sec- 
tion is intended as a guide to provide a basic description of these methods 
for engineers who are unfamiliar x«7ith them. A cost-effectiveness analysis 
of two of these techniques— open-graded asphalt overlays and pavement groov- 
ing is found in Section VTI of this report. A fifth method for restoring 
the macrotexture of existing pavements is the sprinkle treatment. However, 
recent experience with sprinkle treatments has been disappointing due to 
aggregate pickup. For this reason, sprinkle treatments are not included 
in this report. 



A. Open-Graded Asphalt Overlays 

The use of open-graded asphalt pavement surfaces for new pavements 
was discussed in Section IV. A. of this report. Most of the information on 
new open-graded asphalt surfaces discussed in that section is also applicable 
to open-graded asphalt overlays, and will not be repeated here. This section 
focuses on the aspects of open-graded asphalt surfaces that are unique to 
their use on existing pavements. 

1. Specification ; The typical specification for open-graded as- 
phalt surfaces given in Appendix A and the mix design procedure given in 
Appendix B are applicable to overlays as well as new surfaces. 

Proper preparation of the existing surface is vital to assure the 
success of an open-graded overlay. If the existing surface is rough, un- 
even or extremely cracked, a leveling course should be used under the open- 
graded surface course. Such leveling courses will provide a smooth ride 
and help extend the life of the surface course by preventing early failures. 

2. Cost ; The cost estimates discussed in Section IV for open- 
graded asphalt surface courses also apply to open-graded asphalt overlays. 
The average in-place cost of a 25.4 mm (1 in.) overlay is estimated at $1.20 
per sq m ($1.00 per sq yard). In current state practices, the thickness 

of an open-graded asphalt overlay can range from 12.7 to 38.1 mm (1/2 to 
1-1/2 in.)« The Federal Highway Administration recommends a maximum thick- 
ness of 25.4 mm (1 in.) for an open-graded asphalt overlay. Naturally, 



58 



the overaly thickness has important cost implications. If overlays as thin 
as 12.7 mm (1/2 in.) perform adequately, this type of improvement will be 
more attractive economically. More evaluations of such overlays are needed. 
In the meantime, a suggested rule is that an open-graded asphalt overlay 
should be at least twice the thickness of the maximum aggregate size, e.g., 
at least 25.4 mm (1 in.) thick for a mix with a maximum aggregate size of 
12.7 mm (1/2 in.). 

3. Service life ; The discussion of service life for open-graded 
surfaces in Section IV is also applicable to open-graded overlays. There 
are no reliable data to determine the relative service life of open-graded 
and dense-graded surface courses. However, there is currently no strong 
evidence to indicate that the service life of a properly constructed open- 
graded surface is shorter than that of a dense-graded surface, approximately 
10 to 15 years. 

4. Performance ; Performance problems such as raveling and flush- 
ing mentioned for new open-graded asphalt surfaces can also occur with open- 
graded overlays. In addition, reflective cracking of the overlay is a poten- 
tial problem that often occurs more quickly on open-graded overlays than on 
dense-graded overlays, particularly for surfaces placed over portland cement 
concrete. This problem can be alleviated with use of a binder or leveling 
course. 

There has been no comprehensive evaluation of the accident reduc- 
tion effectiveness of open-graded asphalt overlays. Studies from California 
and Virginia provide some indication of the effectiveness. A before-after 
study of 10 sections in California overlayed with open-graded asphalt con- 
crete found that the wet-pavement accident rate was reduced by 70% in the 
after period..iZ' An evaluation of one section in Virginia resurfaced with 
an open-graded asphalt overlay found that 39% of all accidents before the 
overlay occurred on wet pavement, whereas only 17% of all accidents after 
the overlay occurred on wet pavemenr. 49/ However, neither study provides any 
quantitative measure of the improvement in macro texture or skid number or 
any comparison with the accident reduction effectiveness of dense-graded 
overlays. 

5. Advantages and disadvantages ; The advantages and disadvantages 
enumerated in Table 2 apply also to open-graded asphalt overlays. The only 
additional disadvantage to be added for overlays of open-graded asphalt is 
the problem of reflective cracking described above. 



59 



B« Pavement Grooving 

Pavement grooving is the process of making a pattern of parallel, 
shallow cuts of uniform depth, width, and shape in the surface of an exist- 
ing pavement. The grooves are usually cut with a diamond saw. Portland 
cement concrete pavements are most frequently grooved, but grooving has been 
accomplished successfully on older bituminous pavements where the asphalt is 
well-cured* Grooving should not be confused with texturing of new portland 
cement concrete pavements, xdiich is a finishing process that is accomplished 
while the concrete is in the pals tic state. The International Grooving and 
Grinding Association makes the distinction that patterns spaced less than 
12.7 mm (1/2 in.) center-to-center are considered to be "texturing" and 
patterns spaced more than 12.7 mm (1/2 in.) center-to-center are considered 
to be "grooving . "2^/ 

Most grooving in the United States uses center- to-center spacings 
between 12.7 mm (1/2 in.) and 38.1 mm (1-1/2 in.). Both longitudinal and 
transverse grooves have been used, but longitudinal grooves are more common. 
The objective of this treatment is to place grooves in the tire-pavement 
interface which provides a path for water to escape from under the tire. 
Thus, grooving acts like other forms of pavement macrotexture in reducing 
the potential for hydroplaing. However, despite its proven wet-pavement 
accident reduction effectiveness, longitudinal grooving does not normally 
increase the skid number of a pavement surface. Thus, this form of texturing 
does not influence skid number in the same manner as the random macrotexture 
of open-graded asphalt surfaces or the pattern macrotexture of finished 
Portland cement concrete surfaces. There is some indication that transverse 
grooving may increase skid number. 

Some foreign countries have employed pavement grooving effectively 
for a different purpose than used in the United States. Zipkes reports on 
a highway section in Switzerland that has been grooved transversely at center- 
to-center spacings of 254 to 1,016 mm (10 to 40 in.)*2^' This wide spacing 
does not necessarily keep a groove within the tire footprint at all times, 
as with more closely spaced grooves. The only objective of widely spaced 
grooving is to facilitate the flow of water off the pavement to the shoulder. 
The reduction in the depth of water on the pavement reduces the potential 
for hydroplaning by increasing the critical speed at which hydroplaning will 
occur. 

1. Specification ; There are some variations in the depth, width 
and spacing of grooves that have been used in the United States. Typical 
ranges are 2.4 to 6.4 ram (0.095 to 0.25 in.) for groove width, 3.2 to 0.64 
mm (0.125 to 0.025 in.) for groove width, 3.2 to 0.64 mm (0.125 to 0.025 
in.) for groove depth and 12.7 to 38.1 mm (0.5 to 1.5 in.) for center-to- 
center spacing. The most common specification for grooves is 2.4 mm (0.095 
in.) width, 6.4 mm (0.25 in.) deep, and 19.1 ram (0.75 in.) center-to-center 
spacing. The grooves cut by a diamond saw have a rectangular cross-section. 

60 



The usual practice in most states is to groove the center 3 m (10 ft) por- 
tion of a 3«7 m (12 ft) lane and leave a 0#3 m (1 ft) strip ungrooved at 
the edge of each lane* Appendix B contains a sample specification for pave- 
ment grooving used by the Louisiana Department of Highways* 

The depth of pavement grooves is extremely important to their 
proper functioning as a wet-pavement accident countermeasure . Most state 
highway departments have adopted a modified tire tread depth gauge to in- 
spect the depth of grooves on construction projects. This relatively 
simple, but effective, procedure is illustrated by Pennsylvania Department 
of Transportation Test Method No. 629 also discussed in Appendix B. 

2» Cost ; The cost for pavement grooving is quite variable and 
depends on the construction contractor's familiarity with grooving equip- 
ment and the hardness of the aggregate in the pavement surface course* 
Several contractors report typical productivity rates for longitudinal 
grooving of 0*6 lane-km per hour (0*4 lane-miles per hour). However, ex- 
tremely hard aggregates produce a noticeable decrease in productivity rates 
for grooving* 

The best available cost estimate for longitudinal pavement groov- 
ing is $1*20 per sq m ($1*00 per sq yard)* However, lower unit costs for 
grooving are reported in areas where grooving is used extensively* For ex- 
ample, a typical cost for pavement grooving in the Los Angeles area is 
$0*72 per sq m ($0*60 per sq yard), while a cost of $1*20 per sq m ($1*00 
per sq yard) is more common elsexdnere in the State of California* Transverse 
grooving of in-service pavements is more time-consuming and more expensive 
than longitudinal grooving* While longitudinal grooving can be accomplished 
by closing one lane of traffic at a time, transverse grooving requires at 
least two lanes to be closed for equipment moving* One equipment manu- 
facturer reports that longitudinal grooving can be accomplished 50 times as 
fast as transverse grooving* 

3* Service life ; The service life of pavement grooves depends 
on the type of traffic to ^ich they are exposed* High traffic volumes 
shorten the service life of pavement grooves, but this effect has not been 
adequately quantified* The presence of tire chains or studs on vehicles 
in the traffic stream has an important effect on grooving service life* 
California reports that grooves 3.2 mm (1/8 in.) deep on highways where tire 
chains and studs are not used have a service life of 8 to 10 years ,^/ but 
Pennsylvania reports service life of 3 years or less where tire chains and 
studs are used*2' 

4* Performance ; Dramatic reductions in wet-pavement accidents 
have resulted from pavement grooving* Two California studies completed 
in 1972 and 1975 have found reductions in wet-pavement accident rate of 
73 and 70%, -rpgppr.M'vpl y. 39,77 / xhe largest decreases reported were in 



61 



sideswipe, fixed object and rear-end accidents. However, the accident re- 
duction effectiveness of grooving does not appear to be consistent. In the 
1972 California study, 27 projects decreased in total accident rate, while 
11 projects increased. The change in total accident rate with grooving for 
these 38 projects ranged from a 100% reduction to a 45% increase. A review 
of 77 grooving projects in 13 states reported by Rasmus senJSii' showed an over- 
all decrease of 75% in the number of wet-pavement accidents. The before and 
after periods in this evaluation range from 2 months to 5 years in length 
and the decreases in the number of wet-pavement accidents for individual 
projects ranges from 16 to 100%. 

Some users have also observed that longitudinal grooving is ef- 
fective in increasing the directional control of automobiles. Apparently, 
the automobile tires penetrate slightly into the grooves and form a mechani- 

O 1 / 

cal interlock that helps to hold the vehicle in alignment with the roadway ..2_L' 
However, a persistent concern exists about handling difficulties of motor- 
cycles and small cars on grooved pavements. Pavement grooves do produce a 
sensation of instability while riding a motorcycle, but a recent study 
sponsored by the California Department of Transportation in which seven 
motorcycles of different sizes were driven by two riders on grooved pave- 
ments, found no significant control problem.2Z' Furthermore, the 1975 ac- 
cident study in California found decreases in the number of motorcycle ac- 
cidents after grooving on both wet and dry pavements, even though total 
motorcycle registrations and, presumably, motorcycle traffic on the study 
sections increased by 14.5% between the before and after study periods. ■ZZ''^ 
However, because the sensation of instability is unsettling to motorcyclists, 
even though it does not lead to loss of control, the use of warning signs at 
the beginning of grooved pavement sections is recommended. 

5. Advantages and disadvantages : The advantages and disadvantages 
of grooved pavement as a wet-pavement accident countermeasure are summarized 
in Table 6. 

C. Cold Milling 

Cold milling is a technique for retexturing pavements that has 
come into widespread use recently. Major technological advances within the 
past 2 years have produced a new type of rotary milling machine that can 
grind or scarify pavement surfaces more efficiently and accurately than ever 
before. The milling machine used has a rotating drum with carbide steel 
teeth that can texture a pavement surface with a pattern of grooves that 
provide macrotexture for water drainage at the tire-pavement interface. 
Unlike grooving accomplished with a diamond saw, the grooves made by a mill- 
ing machine are short and discontinuous. The depth of the grooves can be 
varied by controlling the forward speed of the milling machine and the 
rotating speed of the drum. The microtexture of the surface between the 
grooves is a function of the type of aggregate used in the pavement. 
Cold milling can be used to retexture both asphalt and portland cement con- 
crete surfaces. 

62 













tu 


















-C 




CO 














■U 




0) 










0) 




tn 




> 










■u 




tu 




CO 










OJ 


• 


u 




rC 










M 


TJ 


3 














o 


0) 


-o 




>^ 










C 


M 


0) 




CO 










o 


3 


U 




e 










o 


O 


tn 




tn 










tn 


1—1 


c 




l-i 










■ D 


1—1 


•H 




CO 










O 


(U 


CO 




CJ 










c 


S 


x: 


• 












•1-1 




o 


4J 


1—1 










e 


en 




c 


r— 1 










D 


•H 


0) 


tU 


CO 










JJ 




!-< 


e 


e 










•H 


4J 


•H 


(U 


tn 








CO 


,0 


I— 1 


•U 


> 










0) 




to 




CO 


U-; 








w 


^4 


x; 


u 


P- 


O 


• 






CO 


o 


CI, 


o 






?-, 






4-1 


U-l 


tn 




TD 


tn 


4J 






c 




CD 


tn 


CU 


u 


•H 






CO 


13 




OJ 


> 


d) 


I— t 






> 


OJ 


0) 


i-) 


O 


> 


•H 






T3 


in 


J 


•H 


o 


•<-\ 


^ 






«0 


3 


XJ 


•u 


!-j 


u 


CO 






tn 








00 


-3 


4J 






•H 


(U 


tn 


T3 






tn 






Q 


^ 


tn 


0) 


14-t 


-3 


c 










OJ 


TJ 


o 


C 


•1-t 








JJ 


1—1 


na 




CO 










O 


C 


3 


q; 




tj-l 








c 


3 


4J 


<4^ 


tn 


o 








c 




tn 


•H 


jj 










CO 


oi 




1—1 


tn 


3 








o 


u 


u-t 




•r-. 


O 










C 


o 


0). 


r— 1 


•H 








60 


OJ 




CJ 


O 


U 








C 


£ 


<u 


•H 


>^ 


CO 








•H 


CU 


tn 


> 


O 


tn 








> 


?> 


3 


u 


J-( 


c 








o 


CO 




<u 


o 


OJ 




O 




o 


Cu 


01 


tn 


4J 


tn 




2 




v< 




J2 




o 






1— 1 




o 




H 




S 






> 


















O 


















o 




* 









« 




vO 






1—1 




CM 




(^ 




cd 


















-1 


H 
















^ 


2 

U3 




1 

J3 












H 


s 




3 














S 




10 




-3 


jj 




3 




> 








<U 


CO 




O 




!< 




<u 




j: 






O 




PL, 




rQ 




tn 


o 




tn 



tn 

OJ 

&a 

CO 
iJ 
3 
CO 
> 

< 



<U 

u 

c 

0) 
•H 
S-i 
tU 

a 

X 

CU 



C -3 
<U ti) 



•H 
O 
tJ 
CO 

4J 

3 

e 
> 

CO 
Cu 
1 
JJ 

<u 



o 
o 

CJ 
CO 

0) 

c 

CO 
CJ 

to 

c 

•H 
> 

o 

o 
(-* 
ao 



tu 

c • 

CO -3 

,— I tu 

tn 

QJ O 

C r-J 

O CJ 

>^ OJ 

—I ^ 



CO 

3 

•H 

•3 

3 

4J 
•H 

00 

C 

o 



•3 
3 



tn 



3 
tu 

e 

> 

CO 



CO CO 



i-H (U 

^ s 

U -H 

•H -U 
3 

cr CO 



ao 

3 

•H 
> 

o 
o 

ao 

u 

0) 



CO CO 

u 

H 



63 



Several state highway departments have recently experimented with 
cold milling to improve skid resistance of pavement surfaces. The Texas De- 
partment of State Highways and Public Transportation has recently milled 
several highway sections using machines from two manufacturers* Milling was 
used successfully on one recent job on 22*5 km (14 mile) portion of the North 
Central Expressway in Dallas* One pass of the milling machine removed a 
38»1 ram (1-1/2 in.) asphalt overlay and retextured the underlying portland 
cement concrete* The Iowa Department of Transortation has retextured two 
Portland cement concrete pavement projects totaling 0*8 km (1/2 mile) in 
length* Gold milling has also been used on highways in Michigan, Ohio, 
Pennsylvania, and Wyoming* 

1* Specifications ; The advances in cold milling technology have 
been so recent that, to our knowledge, no states have yet adopted formal, 
published specifications for the use of milling to increase skid resistance* 
The Texas Department of State Highways and Public Transportation reports 
that milled portland cement concrete surfaces have average texture depths 
of 2 to 2*4 mm (80 to 95 milli-in*), as determined with the sand patch method* 
This is substantially above Texas* specified texture depth for new portland 
cement concrete surfaces, but it should be recognized that comparisons of 
sand patch test results between milled surfaces and finished new surfaces 
may not be appropriate* 

2* Cost ; The only available information on the cost of cold 
milling comes from the recent jobs in Texas* The cost of retexturing is 
$0*48 to $0*54 per sq m ($0,40 to $0,45 per sq yard) and the cost for both 
removing an asphalt overlay and retextuiring is $0,84 to $1,32 per sq m 
($0,70 to $1*10 per sq yard)* Both of these cost estimates include haul- 
ing the removed paving material from the site* Furthermore, the paving 
material removed from the surface may be recycled into new pavements re- 
sulting in additional cost savings* Productivity rates have approached 
16,700 sq m (20,000 sq yards) per day. 

3* Service life ; No pavement surfaces retextured with the new 
milling machines have been in service for longer than about 1 year* There- 
fore, there are no data available to evaluate the service life of milled 
pavement surfaces* 

4* Performance : There are no reports of performance problems 
with milled pavement surface during their initial period of service* The 
Iowa Department of Transportation has reported an increase in skid number 
at 64*4 km/hr (40 mph) from 25 to 57 on one portland cement concrete pave- 
ment and from 39 to 60 on the other ^Z^/ On the second project, skid num- 
bers were measured at several speeds* The skid number-speed gradient of 
the surface was not appreciably changed by milling* (It remained at a 
relatively low level of 0*75*) There are no other skid resistance data 
available for evaluation and the rate of skid resistance degradation after 
milling is unknown* 

64 



Milling has other benefits besides increased skid resistance. Pro- 
filing of the pavement surface by removal of material reduces pavement rough- 
ness and produces a smoother ride. Bridge clearances on controlled-access 
highways are increased rather than reduced as they would be by an overlay. 
However, there are two potential drawbacks to milled surfaces. The indenta- 
tions in the pavement are not continuous, as with sawed grooves, so ponding 
of water might be a problem. Also, the high texture may increase tire noise 
levels. Currently, there are no reliable data on these potential problems, 
but they merit further investigation. 

5. Advantages and disadvantages ; Table 7 summarizes the ad- 
vantages and disadvantages of cold milling to restore texture of existing 
pavement surfaces. 



D« Seal Coats 

Two kinds of seal coats that are frequently used to improve skid 
resistance are described in this section: chip-seal coats and slurry seals. 
Both of these improvements fall in the general category of asphalt surface 
treatments, a term that describes a broad range of asphalt and asphalt- 
aggregate applications usually less than 25 .4 ram (1 in.) thick. These types 
of seal coats are generally used only under very low traffic volume condi- 
tions, but are described here for the sake of completeness. 

1. Specification ; A chip-seal coat consists of a coat of as- 
phalt sprayed on an existing pavement followed by a layer of cover aggregate. 
Beaton—/ states that an advantage of this type of seal coat is that it pro- 
vides the pavement designer a great deal of flexibility, since his choice of 
a maximum aggregate size allows development of gradations that will produce 
a variety of macro texture depths. It is critical, however, that good polish- 
resistant aggregate be used to maximize the life of the chip-seal coat. Most 
state highway department specifications require the use of a crushed aggre- 
gate. The usual thickness of a chip-seal coat is approximately the same as 
the maximum size of aggregate used. The design method used by most agencies 
to determine the required asphalt content for chip-seal coats is presented 

to I 0/1 

by McLeod«--i2.' Callaway and Epp&=--i' provide an excellent discussion of mate- 
rial selection considerations for chip-seal coats including the selection of 
aggregates and gradation requirements. 

A slurry seal, as specified by most agencies, is a form of sand 
asphalt mixture. It consists of a mixture of asphalt emulsion, fine aggre- 
gate, mineral filler, and water. The first slurry seals used an anionic 
emulsion that required 12 to 36 hr to cure before it could be opened to 
traffic, but the newer slurry seals constructed with cationic asphalt emul- 
sions require only 3 to 5 hr, with proper weather conditions, before open- 
ing to traffic.^' Typically, slurry seals are placed 6.4 to 9.5 ram (1/4 to 

65 



4J 

o 

c 

0) 
!-i 
CC 

03 
T5 
•H 
O 
> 



o 

M 

I-:} 



Q 

o 

o 

>-' 
pa 



en 




XI 




• 


<U 




1-1 




<u 


fit 


• 


03 




03 


CO 


c 


03 




CO 


•u 


3 


O 




<a 


c 


O 


a 




u 


m 


C 






o 


> 


^ 


03 




c 


X) 


c 


•H 




•H 


nj 







• 




ca 




U 


T3 


>^ 


•H 


03 


OJ 


<u 


CO 


Q 


•H 


£0 


4J 

O 


e 




CU 


5 


0) 


03 




^^^ 




c 


1—4 




•H 


14-1 


c 


0) 




1—1 


o 


o 


> 

a; 




<u 


fiO 


^1 


1-4 




o 


c 


10) 






•H 


•H 


4J 


<u 




> 


X) 


c 


03 




Vj 


c 


•H 


•1-1 



tu o 



CM 



en 



►J 

PQ 






W 



o 

2: 

M 

H 
X 

H 
M 



cn 

CU 

fit 

CO 
4J 

c 

CO 

<c 




r-H OJ C^ 



66 



3/8 in.) thick* Slurry seals are frequently used to increase skid resistance 
on low volume facilities* However, because this type of seal consists pri- 
marily of fine aggregate, the skid resistance improvement comes from increased 
mi cro t ext ur e • 

2. Cost ; A typical current construction cost for chip-seal coats, 
based on recent experience in Texas and California is $0,42 to $0,60 per sq m 
($0,35 to $0,50 per sq yard), A typical cost for a 6.4 ram (1/4 in.) thick 
slurry seal is $0.36 to $0.54 per sq m ($0,30 to $0.45 per sq yard), based on 
1975 prices in Pennsylvania and New York, This estimate agrees well with an 
estimate of $3,700 per km ($6,000 per mile) for two lanes ($0,51 per sq m 
($0,43 per sq yard)) for slurry seals reported by Runkle and Mahone. — • Costs 
for a double application of slurry seal (9.5 ram (3/S in. thick)) in Pennsylvania 
ranged as high as $1,61 per sq m ($1,35 per sq yard) in 1975. The prices 
quoted in 1975 would undoubtedly be higher today, 

3. Service life ; Pennsylvania has attained up to 4 years service 
from a 6,4 ram (1/4 in.) thick slurry seal when properly constructed at sites 
where the traffic volume is less than 2,000 vehicles per day. Double appli- 
cations of slurry seal (9.5 ram (3/8 in. thick)) are recommended to preserve 
the service life when ADT is greater than 2,500.12/ The service life of chip- 
seal coats is assumed to be similar to the service life of slurry seals. 

4. Performance : Chip-seal coats have been found to perform well 
when properly designed and constructed. Good performance requires proper 
preparation of the surface before placing the chip-seal coat. Binder or 
leveling courses should be placed on the existing surface, if needed to im- 
prove structural strength or rideability. It is important that the exist- 
ing surface or any added binder course be adequately compacted before a chip- 
seal surface is placed so that the aggregate will not penetrate the underly- 
ing surface and lead to flushing. The asphalt content should be determined 
carefully, because flushing will occur if the mixture is too rich while 
raveling will occur if it is too lean. The application rate of cover ag- 
gregate should be minimized, consistent with the desired thickness of the 
chip-seal coat, to avoid problems such as flying chips, and crushing of ag- 
gregate. It is extremely important that clean aggregate be used in seal 
coats, because asphalt will not adhere to dirty aggregate. 

The expected skid-resistance characteristics of chip-seal surfaces 
are not well defined and vary with the microtexture and macrotexture of the 
surface in a manner similar to other pavements. The initial skid number of 
a chip-seal surface could be predicted using British Portable and sand patch 
test results from sample surfaces constructed for a particular choice of ag- 
gregate and gradation. Zube and Skog§2/ suggest an alternative procedure 
for estimating the initial skid number using the same California oil equiva- 
lent test that was described in connection with the design of open-graded 
asphalt surfaces. 

67 



The performance of slurry seals is similar to chip-seal coats. 
Raveling and flying chips do not present a problem, because no coarse ag- 
gregate is used. However, preparation of the underlying surface is ex- 
tremely important. Excessively cracked or deteriorated bases and bases that 
are not structurally sound will cause the slurry seal to perform poorly.^' 
Also, excessive asphalt contents should be avoided to prevent bleeding.^' 

Both chip-seal coats and slurry seals are intended for use under 
low traffic volume conditions, but there is no agreement on an appropriate 
upper limit of traffic volume for their use. For example, the Iowa Depart- 
ment of Transportation recommends the use of chip-seal coats and slurry 
seals only for highways with traffic volumes less than 1,500 vehicles per 
day.—' On the other hand, Texas reports that chip-seals have performed 
well under traffic volumes up to 20,000 vehicles per day, if high-quality 
aggregate is used. 

5. Summary ; No formal cost-effectiveness analysis of chip-seal 
coats or slurry seals is made in this report because there are no reliable 
data on their accident reduction effectiveness and service lifes. However, 
it is clear that chip-seal coats provide a level of macrotexture that con- 
tributes both to increasing skid number and reducing hydroplaning potential. 
Slurry seals increase skid number due to an increase in microtexture alone. 
Both treatments are suitable only for lower traffic volume sites where short- 
term improvements are desired. With these precautions, local experience is 
the best guide for selecting situations where seal coats are appropriate. 



I 



68 



VI. CO ST- EFFECTIVENESS OF PAVEMENT MACROTEXTURE IMPROVEMENTS 



An example of a cost-effectiveness analysis for alternative pave- 
ment macrotexture improvement techniques is presented in this section. The 
purpose of this analysis is to compare the costs and benefits of the avail- 
able techniques in typical situations where they might be applied. The ba- 
sic analysis approach is provided by the benefit-cost model described in 
Volumes II and III of this report. However, because the computerized model 
is intended for application at specific sites, the model logic has been 
adapted and generalized for application to representative sites. The cost- 
effectiveness analysis considers the application of each pavement macrotex- 
ture improvement at sites of two area types (urban/ rural) and three highway 
types (two- lane/multilane, uncontrolled access/multilane, controlled access). 
Two typical ADT levels are analyzed for each combination of area type and 
highway type. 

The cost-effectiveness analysis makes use of the best available 
estimates of the costs and benefits for the following wet-pavement accident 
countermeasures : open-graded asphalt surfaces for new pavements, open- 
graded asphalt overlays, and grooving of existing pavements. For illustra- 
tive purposes, dense-graded overlays with both high and average microtexture 
aggregate are also included in the analysis. The analysis incorporates es- 
timates of countermeasure costs, initial skid resistance improvement and 
skid number degradation with time that are assumed to represent national av- 
erage conditions. Such estimates are not well established and are highly 
dependent on local materials and construction practices. Therefore, the 
analysis results should not be interpreted as being representative of all 
states and regions of the United States, The analysis results indicate 
generally which pavement macrotexture improvement techniques have the great- 
est potential for cost-effective reduction of wet-pavement accidents, but 
should not be applied indiscriminately. The analysis does, however, pro- 
vide a framework within which individual values can be modified by highway 
agencies to obtain results that are consistent with local experience for a 
particular state or region. 

Several pavement macrotexture improvement techniques discussed 
earlier in this report are not included in the cost-effectiveness analysis. 
For example, the accident reduction effectiveness and service life of seal 
coats and milled surfaces are not well enough defined to make intelligent 
estimates for a formal analysis. Also, the texturing of new portland cement 
concrete pavements by wire tining produces an improvement in pavement macro- 
texture at little or no increased cost. Any such "no-cost" improvement can 
be adopted as a standard without reference to formal cost-effectiveness con- 
siderations. 

The remainder of this section presents the general approach to the 
cost-effectiveness analysis, the estimates used for construction cost, ac- 
cident reduction effectiveness, and benefit-cost comparisons of pavement 
macrotexture improvements. The final portion of this section provides an 
interpretation of the analysis results, 

69 



A. General Approach 

The general approach to the cost-effectiveness analysis is to con- 
sider the application of each pavement macrotexture improvement at sites 
representative of each of the 12 cells in the analysis reported in Volume 
I. These cells represent two area types (urban/rural), three highway 
types ( two- lane/multi lane uncontrolled access/mul tilane, controlled access) 
and two ADT categories (under 10,000/over 10,000). The analysis employs 
procedures to estimate both the construction cost and the savings of acci- 
dent costs on an equivalent uniform annual cost basis. The benefit-cost 
ratio is then calculated as the ratio of accident costs savings to construc- 
tion costs. In order to accomplish this, the following assumptions are made 

. Each ADT category is represented by a typical ADT, which does 
not change with time. 

. The pavement at the site is wet 20% of the time. 

• The countermeasures analyzed reduce the wet-pavement accident 
rate, but have no effect on the dry-pavement accident rate. 

. The skid number for open-graded and dense-graded asphalt sur- 
faces are predicted from assumed values of microtexture and 
macrotexture by the Penn State relationship. 

• The change of skid number vrith cumulative traffic passages fol- 
lows the logarithmic relationship used in the computerized 
benefit-cost model. The coefficients of this relationship are 
estimated from the literature. 

• The effect of changes in skid number on wet-pavement accident 
experience can be predicted using the relationships developed 
in Volume I of this report. 

• The effect of pavement grooving on accident experience is a 
70% reduction in wet-pavement accident rate. 

The cost of fatal, injury, and property-damage-only accidents 
used are those developed by NHTSA. The average cost per ac- 
cident reflects the accident severity distribution for each 
area type and highway type. 

. The service life of new pavement surfaces and overlays is 10 

years. The service life of a grooved pavement is 3 years when 

tire chains and studded tires are present and 8 years when they 
are not. 

. The interest rate (minimum attractive rate of return) is 7%. 

70 






These assumptions simplify the computerized benefit-cost model so that use- 
ful results can be obtained with simple manual calculations. This simpli- 
fied approach makes it easy for the interested reader to revise the analy- 
sis results by substituting data applicable to his local area. The results 
of the benefit-cost analysis are quite sensitive to three factors: percent- 
age of wet-pavement time, dry-pavement accident rate and ADT, Typical val- 
ues of these three factors have been assumed in the following analyses. 
Section VI. D. explains how the analysis results can be adjusted for sites 
that deviate from the assumed values. 



B. Construction Cost Estimates 

Construction costs for pavement macro texture improvements are 
highly variable and are dependent on local materials and construction 
practices. The construction cost estimates used for the cost-effectiveness 
analysis are based on averages of recent bid prices reported by several 
state highway departments or obtained from expert opinion of state high- 
way department materials engineers and estimates by the authors. The fol- 
lowing construction cost estimates were used: 

Improvement Cost Per Sg M (Sq Yd) ($) 

Open-graded asphalt surface course 1.20 (1,00) 

Dense-graded asphalt overlay with high micro- 1,20 (1,00) 

texture aggregate 

Dense-graded asphalt overlay with average 0.90 (0.75) 

raicrotexture aggregate 

Pavement grooving 1,20 (1,00) 

The above costs are used directly when no prior decision to re- 
surface the roadway is made. However, when it has already been decided that 
a roadway required resurfacing to improve strength or rideability, an addi- 
tional improvement in pavement macrotexture may also be considered. In this 
case, only the incremental cost of the pavement macrotexture improvement, 
over and above the resurfacing cost, is considered, A highway section where 
a dense-graded asphalt overlay is planned to increase the structural strength 
or rideability of the pavement provides an example of this situation. The 
estimated cost of a dense-graded overlay is $0.90/sq m ($0.75/sq yd). Sup- 
pose that an alternative design using an open-graded asphalt overlay is also 
considered, the total cost of the open-graded asphalt overlay is $1.20/sq m 
($1.00/sq yd), but the incremental cost is only $0,30/sq m ($0.25/sq yd) 
greater than the dense-graded overlay. The use of an open-graded asphalt sur- 
face for a new pavement presents a unique case. Open-graded surface courses 
have no inherent structural strength, so they cannot be included in the de- 
sign pavement thickness. The approach commonly employed is to use an in- 
creased thickness (typically 38 mm (1-1/2 in.)) of base course material is 
less expensive than surface course asphalt. The incremental cost of an open- 

71 



graded surface course for a new pavement is estimated as $0.57/sq m ($0,475/ 
sq yd). A summary of all incremental costs used in the cost-effectiveness 
analysis follows: 



Improvement 

Open-graded asphalt surface 
for new pavement 

Open-graded a^.phalt overlay 



Dense-graded asphalt overlay 
(with high microtexture 
aggregate) 



Base Condition 

Dense-graded asphalt sur- 
face (with average micro- 
texture aggregate) 

Dense-graded asphalt over- 
lay (with average micro- 
texture aggregate) 

Dense-graded asphalt over- 
lay (with average micro- 
texture aggregate) 



Incremental Cost 

Per Sq M (Cost 

Per Sq Yd) 

0.568 (0.475) 
0.30 (0.25) 
0.30 (0.25) 



The unit cost per square yard is used to determine the improvement 

cost per mile of roadway for pavement surface improvements. For example, 

pavement grooving of a highway with two, 3.7 m (12 ft) lanes costs $1.20/sq 
m ($1.00/sq yd) or: 



= $8,760/km 



or 



$1.20 1,000 
sq m km 


-^ X 7.3 


$1.00 ^ 1 sq yd ^ 


5,280 ft 


sq yd 9 sq ft 


1 mile 



X 24 ft = $14,080/mile 



For cost-estimation purposes, a 7.3 m (24 ft) width is used for all two- lane 
highways and a 14.6 m (48 ft) width for all multilane highways. However, be- 
cause open-graded asphalt surfaces should be carried onto the shoulder to al- 
low a continuous drainage path and avoid ponding water on the pavement, the 
cost estimates for such surfaces include an allowance for 1.8 m (6 ft) shoul- 
ders on both sides of the roadway. 

The equivalent uniform annual cost of each macrotexture improve- 
ment is computed by multiplying the total construction cost by capital re- 
covery factor. 



72 



1(1 + i)" 
^^ ~ (1 + i)^ - 1 



where CRF = Capital Recovery Factor, 

i = Decimal interest rate, and 

n = Service life (years). 

A service life of 10 years is used for resurfacing improvements, 8 years 
for pavement grooving on sections not exposed to tire chains or studs and 
3 years for pavement grooving where tire chains or studs are used. The 
interest rate for all analyses is 0.07 or 7% per annum. 

C« Accident Reduction Effectiveness 

The accident reduction effectiveness of each pavement macro tex- 
ture improvement by resurfacing (i.e., not including pavement grooving) is 
determined in six steps, as follows: (1) estimating the microtexture and 
macrotexture of the new pavement surface shortly after installation; (2) 
determining the skid number from the microtexture and macrotexture esti- 
mates using the Penn State relationship (discussed in Section II); (3) 
estimating the rate of degradation of skid number with traffic passages; 
(4) determining the skid number during each year of the improvement ser- 
vice life; (5) calculating the accident rate reduction associated with that 
level of skid number; and finally (6) converting these accident rate reduc- 
tions to a uniform annual basis. This six step procedure will be illustrated 
below by reference to an example. 

Table 8 shows the estimates of initial microtexture and macrotex- 
ture for various pavement surface types. The estimates were developed from 
literature, including the work of Nixon et al« — and the authors judgment. 
The initial skid number at 64.4 kph (40 mph) for each pavement surface is 
calculated from the microtexture and macrotexture values using the Penn 
State relationship, developed by Leu and Henry J-i^ 



SN, 



64.4 



= (-3U0+1.38BPN)e-0-230(64.4)(MD) 



-0.47 



where SN . , , = Skid number at 64.4 kph (40 mph), 
64.4 

BPN = British Portable Number, and 

MD = Mean texture depth (mm), as determined by sand patch 
method (I in. = 25.4 ram). 

73 





^ 


J3--- 


"d 


OJ 


^A 


0) 


J2 


4J 


s 


jz 


to 


3 


<t a 


S 


Z 


• e 


•H 




<j- 


■U 


•n 


vO o 


cn 


•H 


<!• 


Cd 


^ 


4J v^ 




cn 


(0 



vO 






CO 



in 



<Ti 



o 

>3- 



ro 



00 



CO 



Q 




M 




fe^ 




CO 




Q 


CO 


Z 


1— 1 


< 


CO 




>H 


W 


hJ 


PS 


< 


S 


z 


H 


< 


X 




W 


CO 


H 


CO 


O 


w 


«3 


2 


O 


w 


^ 


> 

1— 1 




Eh 




o 


Cd 


u 


cej 


fe 


!=3 


fe 


H 


w 


X! 


1 


W 


H 


H 


CO 


O 


O 


Pi 


CJ 


u 




M 


W 


g 


ffl 




H 


J 




<: 


Z 


M 


M 


H 




M 


Q 


z 


Cd 


1— 1 


CO 




3 


fe 




O 


Pi 




W 


cn 


« 








Z 


M 




H 




CO 




M 







4S 




4-1 




0, 


(U 


0) 


M 


Q 


3 




u 


0) 


X 


>-l 


0) 


3 


4J 


4J 


o 


X 


Vj 


<u 


o 


H 


ra 




s 


C 




(0 




0) 




S 



o 



00 



IT) 
CM 



vO 






vO 



CN 



vO 





h 






















(U 






















-e 






















s 






















3 






















S 




















<u 






















u 


0) 























— -{ 




















4-1 


^ 




















X 


03 




















CU 


4-1 




















4-1 


)-( 


o 




O 




m 




in 






O 





CO 




CO 




vO 




in 






u 


p-l 




















u 




















• 


•H 


A 


















a 


S 


•Si 
•H 
4-1 

•H 

V4 

« 






0) 
4-) 


r-! 






4J 


-3 



•H 
-C 
CO 

c 


•H 
■U 
CO 










CO 


4-1 


<u 




g 


j: 


0) 










00 


•H 


4J 




QJ 


4-1 


M 










OJ 


> 


CO 




^ 


(U 












»-l 




00 




CO 


S 


(U 








a; 


00 


<U 


OJ 


T} 


O4 




4-1 






0) 


a 


oo 


CJ 


!-i 


0) 




J= 


CO 






CJ 


CO 


CO 


CO 


00 


CO 


00 





4-1 




0) 


CO 


M-l 




u^ 


00 


3 


c 


4J 


CO 




a 


<4-4 


)-l 


(U 


U 


CO 


•>-• 


•t-( 


CO 






> 


»-l 


3 


^ 


3 






4J 


a 


c 




H 


3 


cn 


3 


CO 


CU 


(U 


to 




3 






CO 




U 




u 


o 


•H 


•3 


(U 




0) 




4.1 


X 


•u 


3 


CO 


X 


3 


Ph 




CJ 


■u 


rH 


a; 


r-H 


4-1 


>4-( 


<u 


CO 






CO 


.H 


CO 


4J 


CO 


X 


M 




CD 


S 




<4-l 


CO 


A 


o 


^ 


(U 


3 


)>l 









Vj 


A 


a 


u 


o. 


4-1 


CO 





>. 


u 




3 


O. 


CO 


o 


cn 


o 




(4-4 


ja 


<4-4 




CO 


CO 


CO 


•H 


CO 


u 


u 












CO 




£ 




CJ 


u 


0) 


T3 


■3 




4J 




ta 




-3 


•H 


CM 


4-1 


a> 


a; 




C 


T3 


01 


x: 


0) 


a 




CO 


3 


3 




(U 


(U 


-d 


00 


-3 




!-< 


cs 


•H 


•H 




S 


-C3 


CO 


•H 


CO 


CU 


O 


•H 


P 


a 




0) 


CO 


u 


X 


!-i 


00 




iJ 


S 


M 




> 


)-i 


00 




00 


CO 


a 


CD 


(U 


(U 




CO 


60 


! 


x: 


1 


u 


< 


(U 


u 


4J 




Pn 


1 


01 


4J 


0) 


CU 






(U 


OJ 






c 


CO 


•H 


CO 


> 


c 


CO 


Q 









(U 


3 


s 


c 


CO 


M 


CO 










Cli 


(U 




(U 




O 




'**.*^ 


^^*^ 






o 


Q 




o 




3 




col 


^1 



74 



These estimated skid numbers are also shown in Table 8^ The analysis as- 
sumes that the existing pavement surface is x^orn AC or PCC, estimated in 
Table 8 to have a skid number of 31,3. This existing surface is to be 
replaced with one of the alternative improvements shown in Table 8. 

The benefit-cost model described in Volumes II and III uses the 
following form for the change of skid number as a function of cumulative 
traffic passages: 

SN = SNi + Cg ln(T) 

v^ere SN = Pavement skid number at any time, 

SNj = Initial skid number, 

C = Coefficient representing rate of change of skid number with 
traffic wear, and 

T = Cumulative traffic passages per lane/ 10 , 

and subject to the constraints: 

T > 1, 

SN < SNjp when C^ > 0, and 

SN > SNj ^en G^ < 0, 

vdiere SNc = Final or limiting value of skid number. 

The initial skid numbers for alternative surface improvements in Table 9 
are used as values for SN-. The typical skid number for a worn AC or PCC 
surface, 31.3, is used as the limiting value for skid number, SNf. The 
cost-effectiveness analysis is sensitive to C^, the rate of decrease of 
skid number with traffic wear. Rizenburgs et al.,-22./ determined a value 
of -8.3 for Cs for typical asphalt surfaces in Kentucky. However, these 
asphalt surfaces contained predominately limestone aggregate, which is 
highly polish-suscepcible and not suitable for open-graded surfaces or 
other surfaces that require high-micro texture aggregates. Rizenburg 
et al . , found a smaller value of Cg, -4.7, for Kentucky rock asphalt which 
is a porous material with surface voids similar to an cpen-graded surface ^ 
This latter value for the polishing race has been used in this analysis. 

The skid number of the improved pavement surface at the midpoint 
of each year is determined by substituting the initial skid number, SNj^, 
rate of change of skid number, Cg, and the cumulative traffic passages per 
lane into the expression given above. It is assumed that the pavement 



75 



surface that existed before the improvement has reached its limiting value, 
31.3, and does not change with time. The difference between the skid num- 
ber of the iii5)roved pavement at the midpoint of a given year and the skid 
number of the existing pavement is used to determine the accident rate reduc- 
tion for that year. 

The reduction in wet-pavement accident rate for a given change in 
skid number can be predicted from the Phase I analysis results presented in 
Volume I of this report. The following wet-pavement accident rate — skid 
number relationship was reported in Volume I: 



AR = AR + a, SN,- 
o 1 AO 

wtiere AR = Wet-pavement accident rate ( accident s/MVK), 

ARq = Zero- intercept of wet-pavement accident rate (a function 
of highway type, area type, and ADT), 

a-j^ = Slope coefficient (accidents/MVK/SN), and 

SN54,4 = Skid number at 64.4 km/hr (40 mph). 

From the last equation it follows that the effect on wet-pavement accident 
rate of an improvement in skid number is: 

AAR = AR - AR^ = a, (SN - SN^) 
a 1 a b 

'sdiere AAR = Decrease in wet-pavement accident rate (accidents/MVK), 

ARg = ¥et-pavement accident rate after improvement (accidents/MVK), 

ARl = Wet-pavement accident rate before improvement (accidents/MVK), 

SN^ = Skid number at 64.4 kra/hr (40 mph) after improvement, and 

SNv — Skid ntffiiber at 64.4 km/hr (40 mph) before improvement.. 

This relationship is used to determine the accident rate reduction from the 
change in skid number. The slope coefficient, a]^, was found to be -0,0286 
accidents/MVK/SN (-0,046 accidents /MVH/SN) for all highway types, area types 
and ADT levels in the Phase I analysis. This value of the slope coefficient 
is used in the cost-effectiveness analysis. However, the Phase I analysis 
also found that the slope coefficient is sensitive to the level of dry- 
pavement accident rate. This sensitivity has important implications for 

76 



1 



the cost-effectiveness of pavement raacrotexture improvements, but will not 
be introduced until a later point in the analysis. 

The expected wet-pavement accident rate reduction during each 
year of the improvement service life is placed on an equivalent uniform 
annual cost basis for comparison with the improvement construction costs. 
The accident rate reduction for each year could be converted to number of 
accidents reduced or accident cost reduction but, at this point, it is 
most useful to retain them as accident rates. The procedure used is to 
obtain a present worth of wet-pavement accident rate reduction by multi- 
plying accident rate reduction during each year by the appropriate present 
worth factor. The total present worth of wet-pavement accident rate re- 
duction is the sum of the present worth of the wet-pavement accident rate 
reductions for each individual year. This sum multiplied by the appropriate 
capital recovery factor to yield an equivalent uniform wet-pavement acci- 
dent rate reduction." 

Table 9 illustrates the calculation of the equivalent uniform 
annual accident rate reduction. The site chosen for this example is a two- 
lane highway section with average daily traffic of 2,500 vehicles that is 
resurfaced with an open-graded asphalt friction course. Columns 1 through 
3 illustrate the prediction of skid number values for each year of the 
analysis period. Column 2 shows the cumulative number of vehicle passages 
per lane at the midpoint of each year. The corresponding values of skid 
number, decreasing from an initial value of 63.6 to a value of 45.9 in the 
10th year, are shown in Column 3. Column 4 contains the difference between 
each yearly value of skid number and the expected value if the pavement 
were not resurfaced. The wet-pavement accident rate reduction in Column 5 
is determined from the change in skid number in Column 4. The present 
worth factors for each year are given in Column 6. The present worth of 
accident rate reduction for each year, shown in Column 7, is the product of 
the entries in Columns 5 and 6. The total present worth of wet-pavement 
accident rate reduction is 3,94 accidents/MVK (6,34 accident s/MVM), the 
sum of all entries in Column 7, This value is multiplied by the capital 
recovery factor (i = 1%, n = 10 years, CRF = 0.14238) to obtain the equiva- 
lent uniform annual wet-pavement accident rate reduction for an open-graded 
asphalt surface — 0,561 accidents/MVK (0,903 accidents/MVM), 



The equivalent uniform annual accident cost reduction could have been com- 
puted here instead of the equivalent uniform wet-pavement accident rate 
reduction. However, this would not have been advantageous at this point 
because of the dependence of accident costs on area type and highway type 
Instead, the assignment of costs to accidents is made just prior to the 
calculation of the benefit-cost ratios for different pavement macrotex- 
ture improvements. The procedure also makes it convenient for a highway 
engineer to use accident cost figures other than those assumed with a 
minimum amount of recomputation. 



77 



a^ 






§ 



CM 



S 



lO 




o 



J3 

u 

o /-^ 

:2 M c; 
o H 

•U U -r 

tn (x) s_> 

(V 



C m 
S 

> c _ . , 

(0 <u 3 <: 

a* ts -o < 

4-1 O « 

0) O 

:s < 



c 

O M 

w c ol 






;^ CO 
^ > 



0) k 

en CJ 

to JD 

0) S 

V4 3 
M 



Q) a, 

x> M -^ 

i * -^ 

X) \0 o 

•H <^ 

CO 03 



CNJ 




00 


CO 


ON 






CVJ 


00 






(>J 


a^ 


r- 


vO 


m 


m 


•<^ 


<^ 


CO 


t^ 


CO 


• 


• 


• 


• 


• 


• 


• 


• 


• 


• 


• 


1— ( 


o 


O 


o 


o 


o 


o 


o 


o 


o 





vOOOoocvjr^roONO^t'— '-d" 
• •••••••••• 

oooooooooocn 























4J 






















o 






















H 


o 


•4- 


m 


ON 


O 


CO 


t^ 


o 


ON 


en 




<f 


n 


^o 


CM 


en 


vO 


CM 


CM 


cn 


00 




en 


p~ 


.—1 


^O 


I— 1 


■vO 


CN 


CO 


<t 


o 




ON 


CO 


00 


r^ 


1^ 


vO 


nO 


IT) 


lO 


m 





oooooooooo 









ON 


CO 


ON 




CM 


C7N 




CO 


O 


ON 


CO 


00 


r-. 


r^ 


r- 


O 


vO 


• 


• 


• 


• 


• 


• 


• 


• 


• 


• 


.-4 
>-• 


1— 1 


o 


o 


o 


O 


o 


o 


O 


o 



r-ivoo mcMONi^iocncM 

oo^\omio<f<tvt<t<f 

• ••••••••• 

oooooooooo 



• ••••••••• 

QOcooONOor~^£iiOLO>cj' 

CMCMCN r— li— li— (r— ii— li— I.— I 



co^r^--ocNvo<tlO^~o<^ON 



I— I m ON <!• CM 

CO nO U-) lD lO 



O On CO r-~ I — \0 lO 

m vt <t <f vt <f <(■ 



> 




V 



^ — \ 






4 



CO 



II 



•H CU 

'4-4 d 

ns J 
u 



CM > O 



4-J 



ct3 00 
1—1 CO 



O Pn 



oo<t-r-ir^coONincNico^ 
cMoo<tC7NLnoN£)CMr^co 



O CM \0 <-i 



lO O lO ON >d- CO CO 
1— I Csi CM CM CO CO <^ 



p 


'-^ 


4J 


r-< 


V-' 


en 


c 


CO 




^^ 


(U 


•r4 


0) 


m 


en 


4J 


Fr 


01 


0) 


•H 


■H 


> 


w 


n 


H 


v»^ 


CM 


H 



r-4CMCO<l-LO\Or^COONO 



78 



c 


^ ^ 


en 


o -^ 


Cs 


o 3 


CU 


CO Ot 


II 


U-l 




O fi 


V4 


o 


o 


X -H 


4J 


+J 4-1 


o 


u o 


CO 


O 3 


14-4 


S T3 




(U 


:S 


4J H 


^4 


<U CD 


o 


en 4-1 


s 


CU CO 




U M 


+J 


a 


c 


4J 


0) 


r^ a 


en 


CO (U 


0) 


4J TJ 


!-l 


O 


p^ 


H 



The equivalent uniform annual accident reduction for other 
countermeasures, except pavement grooving, is determined in an analogous 
manner . 

The effect of pavement grooving on wet-pavement accident rate is 
handled by a different approach, because pavement grooving is known to re- 
duce wet-pavement accident experience with no improvement in skid number. 
Recent evaluations in California have found that pavement grooving results 
in a 707o reduction in wet-pavement accident rate.ZZ' Therefore, the average 
wet-pavement accident rate for each combination of area type, highway type, 
and ADT level were taken from the Phase I analysis results and a 70% reduc- 
tion in these rates was assumed during each year of the service life of the 
grooving. The presence of tire chains or studs in the traffic stream was 
accounted for by a reduction in the service life of the improvement. 



D. Benefit-Cost Comparisons 

The benefit-cost ratio for investigating pavement macrotexture 
improvements is formed as: 



, ^ ACR 
AGC 

where ACR = Equivalent uniform annual accident cost teduction (dollars) , 

and 

ACC = Equivalent uniform annual construction cost (dollars). 

The method of determining construction costs has already been discussed. 
The annual accident cost reduction is calculated as the product of the an- 
nual number of accidents reduced and the appropriate cost per accident. 

The annual number of wet*paveaent abcidents reduced is deter- 
mined from the accident rate reduction in the following manner: 



_ (ARR)(ADT)(FWET)(365),(L) 
1,000,000 



where AAR = Equivalent uniform annual number of wet-.pavement accidents 

reduced 

ARU = Equivalent uniform annual accident rate reduction (accidents/ 
MVK), 

79 



ADT = Average daily traffic (vehicles/day), 

FEWT = Fraction of time with wet-pavement = 0,20, 

365 = Number of days per year, and 

L = Length of section = 1»6 km = 1 mile* 

The costs per accident used in the analyses are taken from the computerized 
benefit-cost model logic and are shown in Column 9 of the benefit-cost tabu- 
lations. These costs are based on the NHTSA accident costs ($200,700 per 
fatality, $7,300 injury, and $300 per property- damage- only accident involve- 
ment) and a typical distribution of accident severities for each combina- 
tion of highway type and area types. The benefit-cost ratios computed by 
this approach are referred to as unadjusted benefit-cost ratios, because 
three adjustment factors discussed below have not yet been applied. 

The benefit-cost ratios developed for macrotexture improvements 
where no prior decision to resurface has been made are found in Tables 10 
through 14. These tables illustrate: 

• Open-graded asphalt overlay (Table 10), 

• Dense-graded asphalt overlay with high microtexture aggregate 
(Table 11), 

• Dense-graded asphalt overlay with average microtexture aggre- 
gate (Table 12), 

, Pavement grooving (Tables 13 and 14), 

The incremental benefit-cost ratios for pavement macrotexture im- 
provements when a previous decision has been made to build a new pavement or 
resurface an existing pavement are found in Tables 15 through 17, The tables 
illustrate: 

• Open-graded asphalt surface for new pavement (Table 15), 

• Open-graded asphalt overlay (Table 16), 

• Dense-graded asphalt overlay with high microtexture aggregate 
(Table 17). 



80 



4) (U 

o c 

a "^ 

LLJ U 



V 


B 


a 


u 








3 


»H 




a 


in 


.u 


<s 


•w 


■H 


2 


u 


•a 


s 


4> 


e 


M 


•a 


o 


4) 
4J 


ki 


<j 


s 





4) 


3 


•H 


IS 



(11) 

Unadjusted 
B/C Ratio 


0.51 
1.76 

0.58 
I.Li _ 


^ r^ (Nj r^ 
o r^ \0 o^ 

o - o o 


to r^ 

o d 


^ en 
<r X 

d d 


(10) 

Unadjusted 

Annual Accident 

Cost Savings 

($) 


u-l CN 
-J u-l 


CM U1 

in o 


■a- (N 

-;r a:' 


00 c^ 

^ CN 


CC vC -J o 
00 CN en — 
r-- m U-l o 

f^l <■ CN u-l 


(9) 
Cost Per 
Accident 

($) 


00 <o 

O CI 

ON Oi 


SO QO 

M ■ 


o o 

CM CN 

m m 
ON Oi 


CN CN 

m ui 
in ui 


r. r-. (N <M 
a* o\ r- r^ 

00 GO \0 so 

^ -3- u-l UI 


(8) 

Annual 

Accidents 

Reduced 


— in 

o o 


r- r-i 
-T CO 

-3- 00 

O O 


<f 00 

-J- X 

d c 


u-l <o 
en u^ 

d d 


r* r-i 
^ CO 

d d 


r-- en 
<!■ CO 
<t X 

d d 


(7) 
Equivalent Uniform 
Annual Accident 
Rate Reduction 


o 
(J 

'J 
o 

<3 


O -^ 
O O 

in ^ 
O O 


\0 ul 

^ o 
o o 

00 
-4 (^ 

U1 ro 
O O 


vO ul 

-H O 
00 O 

d d 

00 

— * r^ 
u-l m 

d d 


O -1 

d d 

vO en 
»-4 CN 
>* -n 

d d 


sO u-t 
—1 O 
00 vO 

d d 

00 

—1 r^ 
in en 

d d 


vO u-> 

— O 

X vO 

d o 

X 
.-4 r* 
in en 

o o 


(6) 

Equivalent 

Uniform Annual 

Cost ($) 


C O 

o o 




-3- VT 

o d 




sO \0 


o S^ 


(5) 

Construction 
CosL ($) 


O O 

Csl CN 
CN CN 


35 

CN CN 

(N CN 

-J -J 


3 3 

CN CN 

CN CN 

-3- ~T 


O O 
CN CN 

CN CN 


3 3 

CN rM 

^ -J- 


33 

^ C^l 

CN nj 


21 


O O 

o o 

u-l O 

^4 


o o 
o o 
m o 

1^ o 

CM 


O O 
O O 
U1 O 

r-T d 

CM 


O O 

O O 

m o 
r> u-l 


o o 
o o 

u-l O 

* 


o o 

u^ O 
r^ d" 


u 
o 

60 

y~. 4) 

en u 

^ B 
U 

g 
< 


o 
o o 

°.§ 

o » 

-< o 

u 

3) M 
•T5 4) 
S > 

= o 


o 

o o 

o o 

« o 

2o 

u 

41 U 

-3 41 

5^ 


o 

§s 

-J o 

r-l 
ki 

41 b 

■O 41 
S > 

= o 


a 

o o 

o o 

• o 

o • 

-4 o 

Li 

41 U 
•O 41 

S i 
3 O 


o 

o a 
o o 

— o 
u 

41 Ui 

=> o 


o 

o o 
o o 

o°. 

-< o 

<-< 
u 

41 ki 

•a 41 

B > 

s o 


41 

a 

CO 

X 


41 

B 
(0 

hJ 
1 

i 


Multllane 

Uncontrolled 

Access 


Multllane 

Controlled 

Access 


4J 

B 

1 


Multllane 
Uncontrolled 


Multllane 

Controlled 

Access 


(1) 

Area 
Type 


u 

i 


eg 
u 

3 


i-i 
<s 
u 
3 
sS 


B 

m 

3 


a 

OJ 

•a 
u 

= 


B 
a 

J3 
ki 



ki 












a 






kl 


X 


•o 




u 


41 


B 


9) 


£ 


U 


i 


s 




C 


V 


p^ 


u 


41 


•o 





a 


s 




> 




41 


B 




^ 


> 





u 


ki 


9 


■H 




to 


a. 


kl 


<kl 


3 


1 


O 


vu 




u 


IS 


S 


■o 


4) 


kl 


u 


< 


3 


H 


H 



81 



>■ 

o 



(11) 

Unadjusted 
D/C Ratio 


-3- l~- 

o o 


sO ^ in vc 

o o o c 


S3 

o o 


ej> c 

CM -: 

d c 


o o 


(10) 

Unadjusted 

Annual Accident 

Cost Savings 

($) 


eo if\ 


•» - ON 00 

vO P- c ■a' 

00 U- CM O 

^H CM CM r^ 


o^ in 

r- X 


vO c 


m m 
<■ m 

tn 00 


(9) 
Cost Per 
Accident 

($) 


00 oo 


<M CM O O 

NO >0 CNl ts 

00 oo en c 


CM CM f^ IV 

in m CTN o^ 

M -4 00 00 

mm •* ^ 


NO NO 

m m 


(8) 

Annual 

Accidents 

Reduced 


U-l vO 
ON >o 

o — 
o o 


en CN^ en CM 
CM en CM en 

o o o o 


m NO 

O O 


r« e^ 
en CM 
eN .-' 

d d 


en CM 

CM fn 

d d 


(7) 
Equivalent Uniform 
Annual Accident 
Rate Reduction 


u 
o 

U 

o 


u-l rt 

• • 

o o 

eg ov 
en o 

• • 
o o 


(-1 s* 

en CM 

^ CM 

• • 

o o 
o o 

CM p-t 

o o 


en si- 
en ev 

^ CM 

O d 

O O 
r~ -3- 

CM ^ 

o o 


CNJ (N 

CO m 

CM »-( 

• • 

o o 

nO m 
r^ ON 
-1 O 

d d 


<n ^ 
en rsi 

■4* CM 

t • 

o o 
o o 

CM ft 

d d 


tn ~3- 
;n CM 

sj" CNJ 

d d 
o o 

CM -H 

• * 

o o 


(6) 

Equivalent 

Uniform Annual 

Cost ($) 


in m 

88 


o o 


o o 

o_ S 


m m 

c o 
o c 

CM CM 


o o 

o_ S 


s s 


(5) 

Construction 

Cost ($) 


c o 


o o 

vO •£> 

CO CO 
CM CM 


O C 

as 30 

CM CM 


c o 


o o 

00* CO* 
CM CM 


o o 

00 30 
CM CM 


^ < 


§§ 

m o 
CM m 

f-4 


in o 

CM 


o o 
m o 
t«.* O 

CM 


o o 
o o 

m o 
r- m 


O O 
O O 

m o 
r«^ C 

CN 


r<« O 

CM 


U 



^ !f 

< 


O 

-^ o 

1-4 
4) (4 

"O 41 


i§ 

-< o 

(-1 

41 U 
■O 4) 

B > 
3 O 


o 
o o 

-4 o 

b 
4) b 

II 


O 

f-4 

u 

4> U 

II 


a 

a o 
o o 

o^ 

— o 

M 

41 U 
•U 4J 

B > 
3 O 


i§ 

-< O 

^4 

u 

91 b 
"O 41 
B > 

3 O 


4> 

a 

CM a 
-1 

90 

X 


B 

5 
S 


Multilane 

Uncontrolled 

Access 


Multilane 

Controlled 

Access 


4> 

B 

1 


Multilane 

Uncontrolled 

Access 


Multilane 

Controlled 

Access 


(1) 

Area 
Type 


01 


1-4 
a 

b 
3 

as 


1-4 

2 


B 
eg 

14 

3 


B 
eo 
ja 
u 

3 


s 

•2 

3 



3 H E- 



82 



s e 
o o 



e 



e 

V 09 

•3 "O 60 

« ^ C 

.^ 0) o > ^-» 

O 3 < <s <n- 

—I « ^' 

— ' -D -J 

s a u 

e 3 01 

SCO 

c u 

< 



«i a 

^^ OS u ^^ 

o < 



in 

u -o 

^ C 9) 

eg 41 u 

3 -O 3 

e J u 



e -a u 

3 -J o 

/-» o 3 

f". u o -o 

^ a < : 

41 ec 

<e CO 

> 3 

-.J a <o 

3 a BS 

O" < 
Id 



a 
w 3 

e s -^ 
4) a «>■ 

^-\ f-4 ■< *w 

<o a 

v-' > g 4J 

3 

a 



« o 
a cj 
o 
o 



o o 
o 



o o 

o 



CO CO 






o 
ao 

.^ « 

(^ u 

■^ eg 

U 



q 41 
41 c 

< H 



CO 

o o 

• 

o 



o o 
o 



■a- 
o o 






o o 
o 



o o 
o 



o c 
o 



o o 
o o 

in o 



o 

o o 
o o 
• o 
o • 
^ o 



« u 

a > 
= o 



o o 
o o 



o 

o o 
o o 
- o 
o 
-< o 



*' a 

U o 4) 

rH O O 

a a o 

£ 3 < 



o o 
o 



o o 

O 



o o 
o 



o o 
o o 

m O 



O 

o o 

o o 

« o 

o • 

-1 o 



4) ki 

•O 41 

a > 

=3 O 



»j B O 
3 O O 

X u < 



o o 
m tn 



O O 
O O 
in o 



O 

o o 

o o 

• o 

o • 

-< Q 



-3 V 

55 



vj- O 



CM CM 

m m 



o o 
o 



o o 
o 



o o 

o 



■J- 

o o 
o 



o o 
o o, 
m o 



o 

o o 

o o 

• o 
o 

-J o 



a i-i 
a o 

q ki 



o o 
o o 

m o 



o 

o o 

o o 

• o 

o • 



a > 
= o 



-< a 

^5 



o 
a 

4) C 



0) -O 

4) a 
> o 



3 c-> H 



83 



> 
o 
o 
u 
o 





(11) 

Unadjusted 
B/C Ratio 


3^ 

r^ eg 


1 
00 cJ r-» o 
^ ccj irv r-i 


o^ o 

X C7< 

CM r- 


ON vC 
CM C 


m ^ 

O CM 




(10) 

Unadjusted 

Annual Accident 

Cost Savings 

(?) 


-3- cvi 

OJ 


o o o m 

00-3 O CN 

in r- in ^ 

in csj CM ^ 

r-i 


lO t~ 

CM r- 

1 00 ^c 

vO CC 


<f c\ 
NO c 

—1 <y 
en r- 


O ON 

m ON 

en CN 




U 4J 
01 B 

ex. 01 

0^ JJ -w «. 

>-' en o <-■ 
o u 

u < 


00 oo 


CM CN O O 

vo vo rg CN 

00 00 r*l c^ 


c^ CM r». r» 

in in ON oi 

^ -J 00 « 

mm N* ^ 


CM CM 

NO NO 

m m 




(8) 

Annual 

Accidents 

Reduced 


O m 


o i> 

o -J 


00 r- 
■■o m 

CM vO 

o o 


en NO 

* . 
— c-i 


C3N 00 

NO r^ 

CN ON 


-3- 

CM X 

NO CN 

O CN 




(7) 
Equivalent Uniform 
Annual Accident 
Rate Reduction 


o 

O 
OJ 

— , 
u 
o 


iri vo 
O 05 

IN tSl 

CO OV 

• • 

-4 (-4 


CO o 
CM oo 

T-4 CM 

O m 
30 r~ 

O .^ 


o^ in 
■-* -3- 

O O 

• 

o 


CM O 

-3 ^ 

CM en 

00 NO 
CM O 

O <N 


-1 o 

On f^ 

tn <t 


-3- vO 

-( m 

• • 

—1 m 

!>. ON 

• • 
o o 




(6) 

Equivalent 

Uniform Annual 

Cost (?) 


CO 00 

in irv 

n CO 

(N <N 


vT -3 




00 00 

in m 
m en 

O) ONJ 


NO vo 


nO no 

~t -3 




c 
o 

•H ^^ 
U <«• 

/-v O ^^ 

u-i 3 

Wl 

c o 

o 

o 


o o 

00 CO 

o o 


O O 
vO \0 

00 00 
CN CM 


o o 

oo" »* 

CM (N 


o o 

» 00 

o o 


o o 

NO NO 

CO X 
CN CM 


Co NO 

X X 
CM CN 






O O 

o o 

U1 O 

« • 

tvi in 

1-^ 


8§ 

in o 

!>." O 
CM 


o o 
o o 
in o 

r>^ o 

CM 


O O 

o o 

m o 

•ft # 

r«. m 


o o 
o o 
m o 

« 

CM 


in o 

CM 






eo 

^ 0) 

m ij 

^^ CO 

o 
H 

a 
< 


o 
o o 

o 
•> o 

2o 

M 

01 U 
73 01 

B > 
= O 


o 
o o 

o 
• o 

o < 

^ o 

1-1 
u 

01 u 

-a 0) 
c > 

= o 


O O 

*.§ 

o • 
— o 

M 

01 1.1 
■O 0) 

s > 
= o 


o 

o o 
o o 

. o 

So- 

r-4 

5S 


o 

O O 

o 

• G 
o • 
-- o 

u 

01 u 

•O OJ 

a > 

B O 


o 

o o 

o o 

« o 

o - 

'-> o 
<-l 
u 
OJ u 

•O 01 

c > 
B O 




0) 

a. 

CN , CO 

so 

■H 


0) 
B 
CO 

>J 

1 

H 


Multilane 

Uncontrolled 

Access 


Multilane 

Controlled 

Access 


0) 

B 

a 

1 
H 


Multilane 

Uncontrolled 

Access 


Multilane 

Controlled 

Access 




to 0) 

^ OJ o. 
w < H 


a 

M 


2 

3 


1-1 

2 


B 

CO 

M 

B 


B 
<S 

.a 
u 

B 


s 

OJ 

.a 

M 

B 



84 



« W > .-V 

3 < aj <o- 



3 C 





S 


o 


< 




b 


4J 


41 


c 


a, 


<u 


^•s 


•a 


o> U 




-— ' ai 


o 





o 


o 


< 



5 -H 



u 


3 


3 


h 








3 






CO 


m 


■u 


CO 


•^ 


■H 


(U 


o 


T) 


g 


u 


3 


h 


•o 





0) 

■u 


ij 


u 


s 


o 


lU 


3 




Ul 



v^ 3 < O 

CO a 

> 3 iJ 

-" 3 

3 3=: 





3 


3 


,— ^ 




(U 


3 


uv 


*^ 


p^ 


< 


Vk/ 


vO 


n 








> 


id 


u 






u 


to 




3 


o 


o 




cr 


u^ 


CJ 






S 2 



CM CN 

00 oo 



o o 

C^J CM 

en tr\ 



O CO 



n <rl o o 



-3- a 



CM «J 


CO 


•^ 3 


J 


^ 




00 







3 


S 


H 



o o 

O O 



o o 
o - 
m o 



o 

o o o o 
o o o o 

• O - " 

o • 
-I o ^ o 



o o 

• o 

o - 

-- o 



o 
o o 
o o 



- o 
o 
-J o 



o o 
- o 

-. o 



^ iJ to 



•^ C to --^ 



s = < 



E U < 



85 



3 2 



c 

ID 

a. 






M a 



(U) 
Unadjusted 
B/C Ratio 


in in 
— - in 


w^ in 

1 
1 


r-i cNj 
■— m 

CN -^J" 


3- C 


CN r-- 

— CN 

— CN 


ON fn 

CN ^ 


(10) 

Unadjusted 

Annual Accident 

Cost Savings 

($) 


* * 




^ CM 

•J r^ 

C CN 


X o^ 

r- CN 


O nD 
CN CN 
r— en 

CN <" 


CN in 


(9) 
Cost Per 
Accident 

($) 


OS CD 


CN CN 

ca CO 


o o 

CN CN 

en en 
o^ ON 


CN CN 

in in 
in in 


CJN On 
00 X 

« . 


CN CN 

in in 


(8) 

Annual 

Accidents 

Reduced 


X O 

in vo 
^^ in 

O O 


en oc 
vT 00 

O O 


-J M 

CD O 


en in 

O o 


<■ rn 
en X 

-3- X 

o o 


<f en 
en X 
vT X 

c c 


(7) 
Equivalent Uniform 
Annual Accident 
Rate Reduction 


o 
u 

a 

o 

JO 


in r> 

« ^ 

o o 

o ^ 
m r-i 
o o 


.—1 m 

o o 

m CO 

O O 


ON 3 

O O 

in X 
o o 


en r* 

NO -H 

O O 

■4- -n 
— rvi 

<!• m 

o* o 


en m 
ON 3 

O O 

m OT 
ON r^ 
-3- •" 

O C3 


en in 

O 3 

r- vD 

3 3 
m X 

ON l~- 

<t e^ 

3 3 


(6) 

Equivalent 

Uniform Annual 

Cost ($) 


in in 

Ov 3^ 


o o 


o o 
ON e> 


in in 
ON :7< 


On 5 


O 3 

ON o\ 


(5) 
Construction 

Cost (5) 


CO CO 
■X> CO 


lO vO 

m en 
•• . 


CO en 
.-n en 


X X 
X X 
so NO 


vO nD 

r- r- 

n en 

.1 •■ 

en en 


en :n 
en .n 




o o 

o o 

m O 
CM m 


o o 
o o 
m o 

csi 


o o 
o o 

LO O 

r- o 

CN 


o o 
o o 
m o 

r^ m 


O O 

o o 

in c 
r». o 


o o 
o o 

in o 
r~ O 

CN 


U 

o 
ao 

^ D 

^^ eg 

O 

g 

< 


o 

o o 
o o 

« O: 
o •> 

^ o 

kl 
<u u 

•a V 

c > 
=> O 


O 

o o 
o o 

r- O 

u 

<U kl 


o 

o o 
o o 

o°. 
-^ o 

U ki 

•a u 
c > 

= o 


o 

o o 
o o 
« o 
o - 
^ o 

k- 

41 ki 

-g ?: 

3 O 


o 

o o 
o o 

°. 

-> o 

kl 

01 kl 
•0 Hi 

c > 

3 O 


o 

o o 
o o 
« o 
o • 
-■ o 

T— 1 

kl 
eu kl 

•a <u 
c > 
3 O 


4) 
D. 

OI to 
^ 3 

ao 


<u 

e 

CO 

1 
o 

3 


■o 

0) 
U .-1 

c o 

CO b 

^ u tn 

-1 C CO 

u o <u 

.-1 O CJ 

5 - <-' 
Z 3 < 


T3 
V V 

C r-( 

CO ,-< 
f^ to 
-H W CO 

-^ fi o 

3 O O 
S U < 


91 

e 
a 
_] 
1 
o 

3 
H 


-3 
lU 

F^ 

C 
es kl 

-H C en 
u eu 
-4 o o 

S 3 < 


TO 
4) 4) 
C -^ 
CO ^ 
f-4 Q CO 

-H kl en 
ij ij 4> 
-J e o 

i 5 ^ 


(1) 

Area 
Type 


CO 

3 

a: 


1— ( 

!0 
Ul 
3 
OS 


CO 

u 

3 
OS 


C 
CO 

.a 

3 


c 

CO 

.o 
kl 

3 


c 

CO 

kl 

3 



X 


■a 


a 


C 






4J 


P 


c 


OJ 


4J 


T3 



86 



« 



■^ >i 



ill) 

Unadjusted 
B/C Ratio 


0.98 
3.52 


1.14 
2.31 


tn ^ 
CO r^ 


CN <r 




rsi c^ 
X ^ 

o — 


(10) 

Unadjusted 

Annual Accident 

Cost Savings 

($) 


^ CM 

<■ CN 

»— 1 in 


CO m 

-* CT- 


^0 CN 

•<r CO 

O CN 

<f oo" 


O 1^ 
X ON 
r^ CN 


O nC 
CN CN 

r— ' CO 

CN "J" 


CO o 

NO r-l 

<r o 
CN in 


(9) 
Cost Per 
Accident 

(?) 


00 CD 

-3- <r 

<-l c-i 
0\ ON 


CNI CM 
00 00 


o o 

CN CVJ 

CO CO 

cr\ CT\ 


CN CN 

m in 
m in 


r^ r^ CN CN 

ON ON r^ i-» 

X C3D NO nO 

^ vt m m 


(8) 

Annual 

Accidents 

Reduced 


00 M3 

in \ci 

o o 


-* CO 

m CO 

-T CO 

o o 


vT CO 

CO CO 

O O 


CO lO 

CO m 

o o 


vT (O 
CO X 
<f X 


-J CO 

CO X 

<r X 


(7) 
Equivalent Uniform 
Annual Accident 
Rate Reduxtion 


1 

3: 


u-1 r- 
CO m 

o o 

vf CN 

in en 

O 3 


O O 

in 00 

• • 
o o 


-o m 

o^ o 

O O 
tn oo 

<J- CO 

3 O 


o o 

LO CO 

CN csl 

•O CO 


CO m 
ON o 

» • 

m, X 
ON r- 
<t ro 

O O 


CO in 
o o 
r- o 

3 O 

'^ X 

ON r^ 
<f ro 


(6) 

Equivalent 

Uniform Annual 

Cost (?) 


o o 
in in 


o o 
o_ o 

CO CO 


O O 
O O 

CO CO 


o o 
in in 


o o 
c o 

.•o .-o 


CO CO 


(5) 
Construction 


o o 
m m 


CN CN 


o o 

CN CN 
^ cvj 


O O 
in in 

o o 


o o 

CN CN 
CNJ CN 


CN ^1 

r, r, 

CN CN 


<r a 
^ < 


o o 
o o 

u-1 O 
eN4 in 


C O 
O C 
in O 

CN 


o o 

O O 
in o 

P~ o 

(N 


o o 
o o 

in o 
r^ in 


o o 

o o 

in o 

CM 


O O 

o o 

m o 

CN 


u 
o 

CO 

^^ v 
try 4-» 

■^ a 

H 

a 
< 


O 

O O 
O O 

• o 
o • 

r-l O 

u 

(U u 
•3 <U 

a > 
=> o 


o 

o o 
o o 
- o 
o • 
-^ o 

u 

<u u 

c > 

= o 


S o 
o o 
» o 
o • 
— o 

u 
V u 

T3 a 

E > 

= o 


O 

O O 
O O 

- o 
o - 

^ o 

I-I 

01 u 
T3 01 
3 > 

= o 


O 

o o 

O CD 

•> o 

O ' 

— o 

u 

11 u 

Tj y 
c ^ 

=1 o 


o 

o o 

o o 

- o 

o • 

^ o 

u 

v u 
T3 y 
c > 
=> o 


a. 

>^ 

H 

rsl CO 

—' 3 

j: 
eo 


i 

1 

1 


Multilane 
Uncontrolled 

Access 


Multilane 

Controlled 

Access 


y 

c 

CO 

1 
o 

3 

f 


Multilane 

Uncontrolled 

Access 


^tultllane 
Controlled 

Access 


(1) 

Area 
Type 


CS 

u 

3 

ca 


C3 

U 
3 

ta 


CO 

u 

3 
a: 


c 

CO 

.a 

!-i 


c 
to 

Si 

u 

3 


c 
« 

JO 

u 

3 



87 



r- < 







•o 




























(U 





























JJ 




























^v en 


u 


m 


T— J 


o r^ 


C3N 


■J- 


r^ 


f— 4 


o 


o 


r- m 






-J 3 


a 


■~o 


1—4 


r^ m 


o 


O 


m 


r^ 




NO 


CM CO 






t— 1 "I ai 


* 


• 


■ • 


• 


• 


• 


• 


• 


• 


• • 






^^ -o 




I—* 


en 


—1 CN 


CM 


en 


I— t 


r-l 


t— ( 


<— e 


f-4 r-4 






OJ 


U 


























C 


-^ 


























=3 


CO 


























■U 




















































































01 to 


























■a 


•a so 
























v 




























XJ 




r* 


>£) 


in -J 


e-M 


oc 


r^ 


r- 


c:n 


CNJ 


r-) m 




z*^ 01 


o > 
< to 


^^^ 


^^J 


en 


i£) r* 


O 


<• 


CO 


m 


tjN 


O 


r^ in 




O 3 


</> 


CC 


m 


r^ in 


O 


O 


r^ 


00 


o 


NO 


CM 00 










* 


* r 


• 


„'' 




* 




•. M 




■^ -o 








•^ 


f-i CM 


CM 


en 




■"* 


^~* 


t—t r-4 




n 


CO u 
























c 


3 en 
























3 


C O 


























c u 




























< 




























hi u 




























<u c 




CD 


ex5 


^ CN 


o 


o 


CM 


CNJ 


1^ 


r^ 


CM CM 






0, cl> 




<f 


<r 


vD i£> 


CM 


CM 


m 


tn 


ON 


ON 


P^ P*. 




,^ 


•D 


>— s 


en 


fn 


00 OO 


en 


en 


^^ 


_4 


OO 


CO 


NO NO 




ON 


■U 'M 


<n- 


* 


■k 


•t • 


•t 


■^ 


« 


n 






*» «> 






o o 

o < 




ON 


ON 


p~ r~. 


ON 


On 


in 


m 


-* 


<!■ 


in in 






■J] 




























4J 


TJ 


























-1 c 


« 


CO 


vO 


<- r- 


•<r 


r-- 


en 


NO 


<r 


f^ 


vT p- 




*"v 


m <u 


(J 


CO 


^o 


C>J CN 


CM 


1>J 


in 


NO 


CM 


CM 


CM CNj 




33 


3 -O 


3 


o 


< — 1 


CM en 


CM 


en 


(-4 


t— 4 


CM 


en 


cN en 




N-*' 


C -H 


-3 


• 


• 


• • 


• 


• 


• 


• 




• 


• • 






^ 5^ 


& 


o 


c 


o o 


o 


o 


O 


o 


O 


o 


c o 






<: 


























u 
o 




























u c 


en 


e-Ni 


o <■ 


o 


<f 


On 


CM 


o 


-J- 


o ^ 




u-( 


C 


CO 




~* rM 


I — 1 


CM 


r~^ 


in 


1 — 1 


^ 


—4 e^ 




•rH 


u ■-. 


^ 


<f 


—J 


<3- CM 


^ 


CNJ 


CM 


<-^ 


<r 


e^J 


■J- CM 




5 


•3 -u 


■J 
(J 


o 


O 


O CD 


CD 


O 


O 


O 


o 


O 


o o 






O 3 


n 
























P^ u 


-J -o 


























■^ c 


< <U 


























<u 


s: 




























t— 1 


:si 
























CO 


CU dJ 
3 tj 




en 


en 


^ O 


NO 


O 


■<r 


Ln 


NO 


O 


vO o 




z 


y 


o 


C3N 


in vf 


in 


<r 


r~ 


CN 


m 


<r 


m <f 




3 cc 


-^ 


r-5 


O 




CNJ 


>— e 


'—< 


o 


CM 


^ 


C-l r-4 




3 
IT 


3 ca 

< 


O 


O 


O 


O <= 


o 


O 


o 


o 


o 


ID 


o <o 




a 




^ 


























^ 




























ca 




























J-> 3 




























C B 


z— s 


























lU c 


o> 
























\0 




•"^ 
























xj 


^ 


^ 


eM (NJ 


CSJ 


CM 


^ 


^ 


CM 


CSi 


CM CNJ 






05 


c 


o 


O O 


O 


o 


o 


c 


o 


O 


o o 






3 O 





en 


u-1 


o o 


o 


o 


in 


in 


o 


o 


o o 






3- UJ 


o 






























—I 1— 1 


.— ' 


r— I 






>-^ 


f— 4 


r-4 r-4 






c 




























= 




























c 




























o 




























■w 




























4J 


4«- 


























u-1 3 


"^^ 


o 


o 


53 


3 


s 


o 

CM 


o 

CM 


3 


3 


33 






^— ' ^ 


U 


in 


m. 


o o 


o 


o 


m 




o 


o 


o o 






u 


in 




M 










A 












tn 




o 


m 


en 


r-" r-' 


r-* 


r-* 


en 


en 


r^ 


r* 


r- r~ 






5 




























o 
































O 


o 


o o 


o 


O 


O 


a 


O 


O 


O O 










O 


o 


o o 


o 


O 


O 


o 


o 


o 


O O 






z— V 


H 


in 


o 


in o 


in 


o 


in 


o 


in 


C3 


in o 






<f 


Q 








1 




















< 


CM 


vn 


r- O 

CM 


t~~ 


o 

CM 


r-. 


ITi 


r^ 


O 

CN 


p~. O 

CM 








>^ 


o 




o 


O 




O 




O 




o 








U 


o 


O 


o o 


o 


o 


o 


o 


O 


o 


o o 








o 


o 


o 


o o 


o 


o 


o 


o 


O 


o 


o o 








CO 




o 


« o 




o 




o 




o 


. o 






,— .^ 


lU 


o" 


•k 


o • 


o 


• 


o 


■ 


o 




o - 






m 


u 


(—1 


o 


-- o 


^^ 


o 




o 


.-* 


d 


r-l O 








a] 




(—4 


f— e 




r-i 




^^ 




—4 










U 


u 




n 


M 




u 




!-l 




k4 










V 


!-i 


01 u 


<i) 


U 


H) 


h 


eu 


u 


0) u 








H 


T3 


01 


•3 0) 


•T3 


OJ 


■o 


eu 


•3 


0) 


-3 01 








C 


C 


> 


C > 


a 


> 


e 


> 




> 


e > 








< 


3 


o 


=5 O 


=> 


o 


= 


O 


S 


O 


= o 








(a 






•o 










-a 












a. 






(U 










V 












^ 






<— 1 


-3 












■3 












CU -H 


0) 0) 








<U -4 




01 0> 










a 




c o 


c -< 




<u 




C 




c ^ 






,-^ 


5s 


e 




O l-i 


CO -J 




c 




CO M 




CO 1-1 






r>4 


CO 


CO 




^ U 01 


r-^ O 


01 


CO 




i-e JJ 


01 


— 1 O CO 








3 


>j 




•H C en 


•H M 


01 


nJ 




•^ C 


CO 


tJ m en 








X 


t 




u 0) 


4J JJ 


<u 


1 




iJ o 


0) 


jj U 01 








on 







-J o o 


^ c 


o 


o 




^ o 


o 


.-CO 










3 




3 C O 


3 O 


o 


s 




i5 


o 


3 o 








E 


H 




S =J < 


Z O 


< 


H 




< 


s u < 










^ 




^ 


.-e 




a 




c 




a 






to 


(U 


CO 




CO 


CO 




CO 




CO 




CO 






^v (U 


Q. 


'M 




u 


u 




J3 




Xi 




ja 






— 1 IJ 


>> 


3 




3 


3 




M 




u 




;j 




^ < 


H 


03 




03 






3 




3 




= 



o 

a. 



a, AJ u-i 
I u <u 
u eg i« 



The unadjusted benefit-cost ratios for each pavement macrotexture improve- 
ment for each combination of area type, highway type and ADT level are found 
in Column 11 of Tables 10 through 17. For the sake of convenience, all of 
the benefit-cost ratios are stimmarized in Table 18. 

These benefit-cost ratios can be used in two ways. First, in the 
form presented in Table 18, the benefit-cost ratios can be used to evaluate 
the cost-effectiveness of pavement macrotexture improvements under typical 
conditions. In the second use, the typical benefit-cost ratios can also be 
adjusted to be more appropriate to a particular site whose actual conditions 
are known. Three adjustment factors are used: a wet-pavement exposure fac- 
tor; a traction-demand factor; and a traffic volume factor. Tables 10 through 
17 indicate which of these three factors are used in the evaluation of each 
pavement macrotexture improvement. The equivalent uniform annual accident 
cost savings can be converted to apply to specific site conditions by multi- 
plying by the adjustment factors. Because the equivalent uniform annual cost 
savings appear in the numerator of the benefit-cost ratio expression, the 
benefit-cost ratios themselves can be adjusted by direct multiplication of 
the adjustment factors. In all three cases, increases in wet-pavement ex- 
posure, traction demands and traffic volumes above those assumed in the pri- 
mary benefit-cost analysis result in higher benefit-cost ratios, and vice 
versa. 

The first adjustment factor is based on wet-pavement exposure 
time. A typical value of 207. wet-pavement exposure time, was assumed in the 
benefit-cost analysis. For sites which deviate from this typical value, 
the appropriate adjustment factor is the ratio between the actual wet- 
pavement exposure time and the 20% assumed value. A convenient tabulation 
of this factor is found in Table 19. 

The second adjustment factor accounts for traction demand. This 
factor is determined by the dry-pavement accident rate for the site. The 
Phase I analysis results presented in Volume I of this report show that the 
rate of change of wet-pavement accident rate with skid number is sensitive 
to the dry-pavement accident rate. This finding is interpreted to mean that 
highways with high traction demands have both relatively high dry-pavement 
accident rates and an increased sensitivity of wet-pavement accident rate to 
skid number. An adjustment factor based on the dry-pavement accident rate 
is tabulated in Table 20. If the value for dry-pavement accident rate is 
not known, an adjustment factor of 1.00 should be assumed. 

The final adjustment factor is based on the actual at the site 
traffic volume. This factor is calculated as: 



Actual ADT 
Traffic Volume Factor = 



Assumed ADT 
89 



Open -Graded 
Asphalt Surface 
for New 
Pavement 


U1 LO 

in in 
—1 m 


— c<- 

1 


2.13 
4.32 


(^ o 
—1 m 

1 
1 


CN I^ 
—1 CS 

—1 fM 


1.29 
2.63 


c 

> 
o 
o 
u 
o 

XJ 

c 

B 
4) 
> 
CO 


0) 

c 

CD U -D CO 

j: 0/ O: 

CJ -^ -O IJ 

•D TJ -H 

O C 3 H 

^ CD JJ 
■-< M 
H 


u-i vo 
O in 


Cv] o 

m o 
d rn 


CM in 

d d 


CM <}■ 

--I r-1 


CM sd- 


m o 

CO CM 

d -^ 

i 


01 Tl 
J-i 01 0! 
— 1 CO T3 01 

CD -l-l H 
z j: M 


-J CN 

1 '^ 


oc P-1 
—1 vD 


m o 
in m 

d -J 


0-. o 
00 c^ 

(M I-- 


CM O 


in <r 


i) 
o 
to 

t/) 
(J 

OJ 

o 
JJ 

c 
o 

•H 
IT. 
-H 

u 

<s 

u 



u 


-o — ' aj 
Qj >j x: 1-1 
TJ <U 00 s <u 

CO > — JJ 4J 
1- C E X CO 

o a) oo 
1 4J x: ->-> ^ 

0) ^ J-l o u 

cfl a -J M 60 

C J= 3 CJ 60 

cu a -i < 
c cA s: 

< 


in — < 


r~ in 


o o 

CM <n 


r~ l-l 

ITS t^ 

-J l-l 


o o 

r-J VD 

--I l-l 


i 

1 
r^ in 

CM 00 

-- l-l 


1 -D r-J CO 

C 01 CO —1 

a. w a. V 
O >-< CO > 

t; < o 


CO <N 

o^ in 

O fO 


— rn 

I— CN 


—1 CM 


CM ON 


l-l <r 
d -^ 


00 \D 

d I-' 


0) 

o 

tfl 

u-i 
>.' 

M 

o 

C 

1 y 


•H 

O 

z 


CO 
■D ^ 

&> l-l 
-O 0) 

CO t. 

u o 

o 

1 4.1 

01 l-l 
CO CO 

C X 

01 a 

S 05 

< 


C 01 

00 u 

CD 3 01 
U i~i u 

01 X CO 
> O 00 

< 4J OJ 

o u 

■u u 60 


<*■ o 
c; d 


o o 
d d 


o o 


^ O 

o o 
d d 


CM O 

o o 
d d 


CM C 

o o 
d d 


01 

JC u 

6.'! 3 dJ 
— 1 JJ JJ 
S !< CO 
Oi 10 
JZ J^ O 
J-- O iJ 
— 1 tj 60 

3 <J OC 


d d 


-3- vo 

d d 

i 


in \£> 

in r^ 

o d 


o m 
d d 


<?N O 
CM .<t 

d d 

i 


•J- nD 
CO <f 

o d 


1 t) — 1 CO 
C 0/ CO r-1 

a. to c 0) 
O *-! CO > 

c < o 


l-l o 

in f>. 
c -J 

i 


en in 

in rJ 

d -H 


d r^ 


CM r~ 

sC On 

d d 


' — r 

vO CM 

c*^ r^ 
d d 


CM en 
-S- 03 

d d 


H 
< 


o o 
o o 
in o 

rsi in 

1—4 


O O 
C O 
in O 

CM 


O O 
O O 
in O 
.h «. 
r^ O 
CM 


o o 

o o 

in O 

* •> 

r^ in 


o o 
o o 
in o 
•I •. 
t^ o 

CM 


o o 
o o 
m o 

r» O 

CM 


CJ 

CO 

oc 


OJ 

c 
n 

1 


0) 

0) —1 
C p 
CO s-> 

u O 0) 
.-1 o o 
SCO 

s: = < 


■a 

01 01 

C -1 

CD .-1 
.-< O CO 

•H p m 

■U iJ 01 
—1 C u 
3 O 

s o < 


C 

CD 

-J 

O 


T3 

OJ 

^-1 

11 f-l 

C 

CD U 

-J JJ 01 i 
•-I C CO 1 
U 01 i 
--I u o 
3 C O 
S 3 < 

1 


•V 

OJ 01 

a l-l 

CO l-l 

l-l m 
— 1 1- m 

J-l 4-1 Oi 

^ C <J 

a o u 
£ u < 


CO 01 


1— 1 

CD 


CD 
l-l 

i 


CO 


B C j C 1 
CD CD CO 
J3 XI XI 
P )-i IM 



90 



TABLE 


19 












WET-PAVEMENT EXPOSURE TIME 


ADJUSTMENT FACTO 


R 




Wet-Pavement 




Wet- 


-Pavement 






Exposure Time (%) Adjustment Factor 


Exposure Time 


(%). 


Ad.-j 


ustment Factor 


0.00 














1 0.05 






21 






1.05 


2 0.10 






22 






1.10 


3 0.15 






23 






1.15 


4 0.20 






24 






1.20 


5 0.25 






25 






1.25 


6 0.30 






26 






1.30 


7 0.35 






27 






1.35 


8 0.40 






28 






1.40 


9 0.45 






29 






1.45 


10 0.50 






30 






1.50 


11 0.55 






31 






1.55 


12 0.60 






32 






1.60 


13 0.65 






33 






1.65 


14 0.70 






34 






1.70 


15 0.75 






35 






1.75 


16 0.80 






36 






1.80 


17 0.85 






37 






1.85 


18 0.90 






38 






1.90 


19 0.95 






39 






1.95 


20 1.00 






40 






2.00 



91 



TABLE 20 




TRACTION DEMAND ADJUSTMENT 


FACTOR 


Dry-Pavement Accident Rate 




(accidents /MVK) (accidents/ MVM) 


Adjustment Factor 


under 0.67 under 1.082 


0.00 


0.68 1.10 


0.02 


0.75 1.20 


0.11 


0.81 1.30 


0.20 


0.87 1.40 


0.29 


0.93 1.50 


0.39 


0.99 1.60 


0.48 


1.06 1.70 


0.57 


1.12 1.80 


0.67 


1.18 1.90 


0.76 


1.24 2.00 


0.85 


1.30 2.10 


0.94 


1.37 2.20 


1.04 


1.43 2.30 


1.13 


1.49 2.40 


1.22 


1.55 2.50 


1.31 


1.62 2.60 


1.41 


1.68 2.70 


1.50 


1.74 2.80 


1.59 


1.80 2.90 


1.68 


1.86 3.00 


1.78 


over 1.88 over 3.02 


1.79 



92 



The value of assumed ADT appropriate for the site under consideration should 
be selected from Column 4 of Tables 10 through 17. This adjustment factor 
is only approximate, since it accounts for the influence of traffic volume 
on vehicles-kilometers of exposure to wet-pavement, but not the influence 
of increased or decreased traffic passages on skid number. If the actual 
ADT differs substantially from the assumed ADT, then this adjustment fac- 
tor should not be used and the calculation of equivalent uniform annual 
accident cost savings should be repeated using the actual ADT. 

A simple example illustrates the use of the three adjustment fac- 
tors. Consider a rural, two-lane highway section with average daily traffic 
of 5,000 vehicles, wet-pavement exposure time of 257o and dry-pavement acci- 
dent rate of 1.55 accidents per million dry-pavement vehicle-kilometers (2,50 
accidents per million dry-pavement vehicle-miles). Let us assume that an 
open-graded asphalt overlay is considered for this section as a wet-pavement 
accident count erraea sure. No prior decision to resurface the pavement has 
been made. Table 18 indicates that the benefit-cost ratio for an open-graded 
asphalt overlay on a rural, two- lane highway with traffic volume less than 
10,000 vehicles per day is 0,51, Since this benefit-cost ratio is less than 
1,0, it does not initially appear that the overlay is justified. However, 
the tabulated benefit-cost ratio is only appropriate for sites with traffic 
volumes of 2,500 vehicles per day, wet-pavement exposure time of 207o and av- 
erage dry-pavement accident rate. The appropriate wet-pavement time adjust- 
ment factor from Table 19 is 1,25; the traction demand adjustment factor is 
1,31, The traffic volume adjustment factor is: 



Actual ADT _ 5,000 _ 2 n 
Assumed ADT ~ 2,500 ~ ' 

The adjusted benefit-cost ratio is: 

0.51(1.25)(1.31)(2.00) = 1.67 > 1.0 



Thus, an open-graded asphalt overlay for this site is economically justified, 
even though it would not be justified for a typical rural, two-lane site. 



E. Interpretation of Benefit-Cost Ratios 

The benefit-cost ratios presented in the previous section after 
appropriate adjustment factors are applied, provide an indication of whether 
a pavement macrotexture improvement is economically justified under general- 
ized site conditions. Improvements whose benefits exceed costs have benefit- 
cost ratios greater than 1.0 and are economically justified. Conversely, 

93 



I 



improvements with benefit-cost ratios less than 1,0 are not justified eco- 
nomically. When several alternative pavement raacrotexture improvements are 
considered at a particular site the final selection should be based on an 
incremental benefit-cost analysis. For an incremental analysis, the alter- 
natives must be arranged in order of increasing costs. Each additional ex- 
penditure (incremental cost) must be exceeded by a corresponding incremental 
benefit in order to justify a higher-cost improvement. This section dis- 
cusses the significance of the benefit-cost ratios for each pavement macro- 
texture improvement. The use of the adjustment factors to establish cost- 
effectiveness warrants for pavement macrotexture improvements is presented 
in Section VII. 

Based on the assumption about costs and benefits in this example, 
open-graded asphalt overlays cannot generally be justified economically un- 
der the assumed conditions except for rural highways in the high ADT cate- 
gory. However, benefit-cost ratios over 1,0 could be attained for any type 
of site with high wet-pavement exposure or traction demand. Also, the com- 
parison of open-graded and dense-graded overlays shows that the use of open- 
graded asphalt is easily justified when a prior decision to resurface a sec- 
tion has been made for reasons other than the reduction of wet-pavement ac- 
cidents. Similarly, open-graded asphalt surfaces for new pavements are also 
cost-effective under typical conditions. 

In this example, dense-graded asphalt surfaces with high microtex- 
ture aggregate do not appear cost-effective as a wet-pavement accident counter- 
measure unless extremely high adjustment factors can be applied or unless a 
previous decision to resurface has been made for other reasons. However, 
even under conditions where dense-graded asphalt surfaces with high-microtexture 
aggregate are justified, the open-graded asphalt surfaces considered in this 
example are more cost-effective, A substantial change in the construction 
costs or rate of change of skid number used in this analysis would be required 
to reverse this general conclusion. Dense-graded asphalt surfaces with average 
microtexture aggregate do not appear cost-effective as a wet-pavement accident 
count ermeasure, under any of the conditions considered in this example. 

The benefit-cost analysis demonstrates that pavement grooving is 
a very cost-effective countermeasure when there are no tire chains or studs 
in the traffic stream, and is cost-effective in many situations even when tire 
chains and studs are present. Furthermore, this analysis probably overstates 
the cost of grooving because nearly the same accident reduction might re- 
sult from grooving only those areas with high traction demands, such as 
horizontal curves. 

The variety of factors that must be considered in the cost- 
effectiveness analysis indicate that there can be no single level of macro- 
texture and no one high-raacrotexture surfacing material that is optimal for 
all site conditions and all geographic areas. It is recommended that the 



94 



selection of pavement macrotexture improvements for a particular site or a 
particular geographic area be based on cost-effectiveness considerations 
such as those presented in this section. The most reliable analysis results 
will be obtained if users incorporate actual construction costs, traffic 
wear rates, wet-pavement exposure estimates, etc«, that are valid for par- 
ticular locations and local materials « Benefit-cost analysis provides a 
strong justification for pavement macrotexture improvements at sites where 
they are appropriate, and can be used to establish priorities between im- 
provement sites. However, it is important in establishing a pavement sur- 
face improvement program that agencies also consider the relative cost- 
effectiveness of other traffic safety improvements not related to pavement 
macrotexture, such as signing, geometric improvements, and roadside safety. 



95 



VII. WARRANTS FOR HIGH- MACRO TEXTURE SURFACE COURSES 



This section discusses the subject of warrants for high-macrotex- 
ture improvements. The initial portion of this section discusses the current 
practices employed by state highway departments to implement pavement macro- 
texture improvements, while the final portion provides an example of the de- 
velopment of warrants from the results of the cost-effectiveness analysis 
presented in Section VI. Because the cost-effectiveness of pavement macro- 
texture improvements depends on wet-pavement exposure time, traction demand, 
and traffic volume, no single set of warrants can be appropriate for all con- 
ditions. The final portion of this section provides guidance for a user to 
develop cost-effective warrants for high-macrotexture surface courses that 
are appropriate for a particular locality. 



A. Current State Practices 

Most states have not established formal warrants for pavement macro- 
texture improvements or other wet-pavement accident countermeasures. However, 
most states have established procedures for monitoring wet-pavement accident 
experience, identifying locations with higher than expected wet-pavement ac- 
cident experience and implementing countermeasures at such sites. The imple- 
mentation decisions in some states are based on some form of cost-effectiveness 
analysis, but in most states countermeasures are selected on the basis of 
engineering judgment, tempered by budget constraints. To a great extent, judg- 
ment has been the only basis on which such decisions could be based because of 
the dearth of proven evaluations of accident reduction effectiveness. In 
nearly every case where cost-effectiveness analyses have been employed, their 
purpose has been to justify the need for some improvement rather than to com- 
pare alternative improvements. Other criteria employed in the decisionmaking 
process include skid number and traffic volume levels. 

Three states that have employed cost-effectiveness considerations in 
the selection of pavement surface improvements are Virginia, California, and 
Texas. Virginia's approach has been described by Runkle and Mahone .-iH' In 
this approach, the annual accident reduction benefit is determined by assum- 
ing that the effect of any surface improvement is to reduce the percentage of 
all accidents that occur on wet-pavement from its level before improvement of 
a particular highway section to the statewide average--20%. Property-damage 
accident costs are estimated as the average of the property damage costs re- 
ported for the analysis site; injury costs are based on the National Safety 
Council estimate of $4,000 per injury; and, fatal accident costs are considered 
equal to injury costs. The ratio of the total constiruction cost to the annual 
accident cost savings, known as the "breakeven value" or payback period, is 
used as one basis for establishing improvement priorities. The Virginia ap- 
proach can only be used to establish the need for an improvement and not to 



96 



compare alternatives, because alternative improvements have the same estimated 
Vaimtral benefit and differ only in the cost. This approach does not incorporate 
the concept of an interest rate or minimum attractive rate of return; does not 
consider the service life of an improvement; and assumes implicitly that the 
in5)rovement is equally effective throughout its service life. Additional 
factors considered by Virginia in establishing priorities for field review 
(including skid testing) and subsequent improvement are number of wet-pavement 
accidents, percentage of wet-pavement accidents and skid number level. 

The California Department of Transportation has developed a cost- 
effectiveness criterion for pavement surface . improvements (and other types of 
safety improvements) called the Safety Index*i2/ The Safety Index is equal 
to the conventional benefit-cost ratio multiplied by 100, except that the 
computation does not involve an interest rate. However, as in Virginia, 
this technique is not employed to compare improvement alternatives. There 
is a surprising reluctance, even by states that choose a formal benefit-cost 
approach, to require a minimum rate of return from improvements through use 
of an interest rate in the analysis. 

Texas has established a priority system for ranking pavement surface 
improvements based on their Skid-Prone Index,^' which they define as: 



/ APT \ 

U,ooo/ 



SPI = B/C X I , ^^^ 1 x SNF 
where SPI = Skid-Prone Index, 

B/G = Benefit-cost ratio, 

ADT = Average daily traffic (vehicles/day) , and 
SNF = Skid number factor = 
= "Iq if SN40 < 30 
= "1^ if 30 ^ SN40 ^ 40 

= i5^^^^A0>^- 

The benefit-cost ratio in the Skid-Prone Index is determined in the conven- 
tional manner from estimated accident cost savings and construction costs 
using an interest rate of 87o. This technique is used to establish improvement 
priority rankings, but not to compare improvement alternatives. It is not at 
all apparent why the ADT or skid number factor should receive independent con- 
sideration in the establishment of priority rankings, if the influence of ADT 
and skid number on wet-pavement accident reduction is incorporated in the 
benefit-cost ratio on a rational and consistent basis. 



97 



Other states have established formal or informal priorities from 
their analysis of locations with high wet-pavement accident experience on 
some basis other than a cost-effectiveness analysis. The most common criteria 
are based on wet-pavement accident frequency or ratio of wet-pavement acci- 
dents to total accidents. 

Other factors that have been used to establish improvement criteria 
are skid number and traffic volume level. Several states have established 
surface improvement guidelines based on skid number measurements, without 
establishing a formal, minimum skid number requirement. For example, one 
state highway department requires continued surveillance of pavements that 
are tested and found to have skid numbers between 35 and 49; corrective ac- 
tion is scheduled in the next fiscal year for pavements with skid numbers 
between 31 and 34; and, immediate resurfacing is required for pavements with 
measured skid numbers less than 30. This kind of guideline establishes the 
need for a surface improvement, but does not indicate which of the available 
alternatives should be selected. The Iowa Department of Transportation,2i' 
has established usage limits for various surfacing alternatives that serve 
as traffic volume warrants. For example, chip-seal coats and slurry seals 
are recommended only for traffic volumes less than 1,500 vehicles per day, 
conventional dense-graded overlays for traffic volumes less than 3,000 
vehicles per day and dense-graded overlays with polish-resistant aggregate 
for traffic volumes from 1,500 to 5,000 vehicles per day. However, no traf- 
fic volume limits have been established for open-graded surfaces, sprinkle 
treatments, and pavement grooving. Such usage limits prevent the applica- 
tion of a surface improvement at an inappropriate traffic volume level. How- 
ever, traffic volume alone is not sufficient to warrant a pavement surface 
improvement. Agencies that have established traffic volume usage limits also 
consider other factors in selecting pavement surface improvements. 



B. Development of Cost-Ef f ective Warrants for Use of 
High-Macrotexture Surface Courses 

The cost-effectiveness analysis presented in Section VI can be 
utilized to develop warrants for pavement macrotexture improvements. War- 
rants alone cannot guarantee that the optimum pavement macrotexture improve- 
ment is selected for a particular site — that can only be accomplished through 
an incremental benefit-cost analysis. However, the warrants can assure the 
user that each pavement macrotexture improvement that is implemented is eco- 
nomically justified. 

The central principle in the development of cost-effective warrants 
is that pavement macrotexture improvements with benefit-cost ratios greater 
than 1.0 are warranted and improvements with benefit-cost ratios less than 
1.0 are not warranted. Three adjustment factors presented in Section VI are 
applicable to sites whose wet-pavement exposure time, dry-pavement accident 
rate and ADT deviate from assumed typical conditions. The boundary between 

98 



I 



site conditions that warrant an improvement and site conditions that do not 
is established by defining combinations of the adjustment factors that pro- 
duce a benefit-cost ratio of exactly 1.0. The warrants developed in this man- 
ner are presented in a convenient graphical form that can be used as a deci- 
sionmaking tool • 

The three adjustment factors are applied to the benefit-cost ratio 
in the following manner: 

Adjusted B/C = (B/C) (WPF) (TDF) (—^^^^Mt^^ 

^Assumed ADT/ 

where B/C = Unadjusted benefit-cost ratio, 

WPF = Wet-pavement exposure time factor (see Table 19), 

TDF = Traction demand factor (determined from dry-pavement accident 
rate; see Table 20), 

Actual ADT = Average daily traffic at site (vehicles/day), and 

Assumed ADT = Assumed average daily traffic for site conditions 
(vehicles/day) (see Tables 10 to 17). 

The influence of the wet-pavement exposure time and traction demand factors 
is completely accounted for by multiplying the unadjusted benefit-cost ratio 
by the factors shown above. However, the influence of ADT is only accounted 
for approximately by multiplying by the ratio of actual ADT to assumed ADT 
as illustrated above. This occurs because the traffic volume also affects 
the rate of change of skid number during the analysis period. To provide ac- 
curate warrants, the unadjusted benefit-cost ratios have been recoirqputed for 
each .ADT level considered using the appropriate rate of change of skid number. 

Warrants have been developed for illustrative purposes for open- 
graded asphalt overlays under two types of site conditions. The two site 
conditions considered are rural, two-lane highways with traffic volumes less 
than 10,000 vehicles per day and rural, multilane, controlled-access high- 
ways with traffic volumes greater than 10,000 vehicles per day. Similar 
warrants could be developed for each of the 96 combinations of area type, 
highway type, ADT level, and pavement macrotexture improvement type shown 
in Table 18. Also, the warrants presented here can be easily modified to 
reflect local experience in areas such as construction costs. Only warrants 
that are consistent with local conditions and experience are likely to be 
used. 

Figure 5 presents the warrants for open-graded asphalt overlays on 
rural, two-lane highways and Figure 6 presents the warrants for rural, multi- 
lane, controlled access highways. The abcissa of each figure represents the 

99 



2.5 



2.0 







> 


i/r 




\ 






«v 


15 






u 

< ] 


.5 


o 

- < 






« 


o 






0£ 






■4- 






c 




1^ 


0) 




c 


-5 




V 






15 


u 






u 




o 


< 




^ 


^_ 






c 




4« 


(U 




c 


£ 
> 




E 

0) 


o 




> 


a: 




£ 



.0 



4.0r- 



3.0 



2.0 



— >s 

Q 



1.0 - 



RURAL. TWO- LANE HIGHWAY 
ADT< 10,000 



OPEN -GRADED ASPHALT OVERLAY 
WARRANTED 




OPEN- GRADED ASPHALT OVERLAY 
NOT WARRANTED 



1 



10 20 30 

Wet Pavement Exposure Time (Percent) 



40 



Figure 5 - Warrants for Open-Graded Asphalt Overlays 

100 



2.5 



4.0 



2.0 



2* 




:S 


> 




> 


^ 




^ 


c 




> 


(U 

-a 




0) 

-a 


u 

< 


1.5 


- ^ 


Sn.^ 






v 




4) 

•4- 


o 




o 


c 




c 


u 

< 






c 




c 


0) 

E 
> 




1 


u 

Q 


1.0 


_ X 

Q 



3.0 



2.0 



1.0 - 



RURAL. MULTILANE, CONTROLLED ACCESS 

HIGHWAY 

ADT> 10,000 



OPEN -GRADED ASPHALT OVERLAY 
WARRANTED 




OPEN- GRADED ASPHALT OVERLAY 
NOT WARRANTED 



10 20 30 

Wet Pavement Exposure Time (Percent) 



40 



Figure 6 - Warrants for Open-Graded Asphalt Overlays 

101 



percentage of wet-pavement exposure time. The ordinate represents the dry- 
pavement accident rate. The plotted curves for different ADT levels repre- 
sent conditions where the adjusted benefit-cost ratio is equal to 1.0, Be- 
cause of the manner in which the traction demand adjustment factor is defined 
(see Table 20), these curves become vertical at dry-pavement accident rates 
above 1.88 accidents/MVK (3.02 accidents/MVM) and approach an asymptote at 
the dry-pavement accident rate of 0.67 accidents/MVK (1.08 accidents/MVM). 

The figures are interpreted in the following manner. The point de- 
fined by the wet-pavement exposure time and dry-pavement accident rate for a 
particular site is located in the figure. If this point lies on or to the 
right of the curve defined for the ADT at the site, the pavement raacrotexture 
improvement is warranted. However, if this point is to the left of the ap- 
propriate ADT curve, the pavement macrotexture improvement is not warranted. 
If no curve is given for a specific ADT level an approximate interpolation 
should be made. 

The procedures presented in this report reflect the current state 
of the art of pavement macrotexture improvement evaluation and analysis. It 
is emphasized that no single set of warrants for high-macrotexture surface 
courses can be appropriate for all circumstances. However, the cost-effec- 
tiveness procedures can be applied to develop warrants applicable to particu- 
lar geographic areas and particular materials. As further evaluations of 
high-macrotexture surface courses are performed, permitting refinements of 
the benefit-cost analyses, the warrants should be updated to reflect the cur- 
rent state of knowledge. 



102 



VIII. CONCLUSIONS 



The following findings and conclusions were drawn from the inves- 
tigations reported in this volume: 

1. Pavement microtexture is determined primarily by aggregate 
properties; pavement macrotexture is determined primarily by mix design and 
aggregate gradation. 

2. Microtexture determines the skid resistance at low speeds. 
Both microtexture and macrotexture contribute to the skid resistance at 
high speeds. Macrotexture is particularly important because it provides 
a channel for water to escape from the tire-pavement interface. 

3. Skid number can be predicted from pavement macrotexture and 
microtexture using a relationship developed recently by Penn State Uni- 
versity. 

4. Numerous methods have been applied by agencies in the United 
States and abroad to measure pavement macrotexture. These methods are de- 
scribed in the text of the report. 

5. The sand patch method is the most widely used and accepted 
method in the United States for measuring pavement macrotexture. 

6. There has been no organized effort to appraise the precision, 
accuracy and intercorrelation of the various macrotexture measurement 
methods. The literature provided only fragmented information available from 
obscure sources. 

7. Many researchers have attempted to validate pavement macro- 
texture measurement techniques by comparison with pavement friction coef- 
ficients (skid numbers). The results of such efforts are bound to be dis- 
appointing because skid number is a function of both macrotexture and micro- 
texture. 

8. There are two primary methods for providing a high level of 
macrotexture in new pavements: open-graded asphalt surfaces and texturing 
of Portland cement concrete surfaces. 

9. Open-graded asphalt surfaces have had wide use in some states 
and very limited use in others. The usual reasons for limited use of open- 
graded asphalt surfaces are high cost and lack of suitable aggregates. 



103 



10. Wire tining has achieved nearly universal acceptance as the 
best method for finishing portland cement concrete surfaces in the plastic 
state. 

11. The methods for restoring a high level of macro texture in 
existing pavements include: open-graded asphalt overlays, grooving, cold 
milling and seal coats. Open-graded asphalt overlays and pavement groov- 
ing are vridely used in some areas. Cold milling has had only limited use 
as a technique for improving skid resistance. Seal coats are appropriate 
only for highways with low traffic volume levels. 

12. Tabulations of benefit-cost ratios for various pavement 
raacrotexture inqirovements have been developed and are presented as an ex- 
ample. 

13. Factors such as traffic volume, wet-pavement exposure time 
and traction demand are often critical in determining whether or not a 
pavement macrotexture improvement is cost-effective. For this reason, no 
single set of warrants for pavement macrotexture improvements can be ap- 
propriate for all site conditions and all geographic areas. 

14. The development of cost-effective warrants for selected 
improvements from the results of the benefit-cost analysis has been il- 
lustrated. The warrants are determined from area type (urban and rural), 
highway type, traffic volume, wet-pavement exposure time and traction de- 
mand (represented by dry pavement accident rate). Warrants developed in 
this manner are applicable to particular site conditions and particular 
geographic areas. 



104 



IX. RECOMMENDATIONS 



The following recoinmendations were developed in the investigation 
of criteria for improvement of pavement macrotexture; 

1. ■ A standard method for macrotexture measurement should be es- 
tablished and applied consistently throughout the highway community. The 
establishment of a standard will require close cooperation between FHWA, 
state highway departments, ASTM and other interested organizations. 

2. More complete information is needed on the relationship be- 
tween pavement macrotexture and wet-pavement accident rate. The indirect 
relationships used in this study should be refined as new information be- 
comes available. 

3. Evaluations of the performance of pavement macrotexture im- 
provement methods should continue. Additional data on the sei-vice lives of 
pavement macrotexture improvements under various traffic conditions would 
be valuable. 

4. Highway agencies are encouraged to select pavement macrotex- 
ture improvements on a cost-effectiveness basis using techniques such as 
those presented in this report. The benefit-cost ratios and warrants pre- 
sented here will be most accurate and most accepted if they are refined by 
individual users to reflect local experience with construction costs and 
paving materials. 



105 



1 



X. REFERENCES 



1. Adam, V., and S. C. Shah, "Evaluation of Open-Graded Plant-Mix Seal 

Surfaces for Correction of Slippery Pavements," Transportation Re - 
search Record , No. 523, Transportation Research Board, Washington 
(1974), 

2. Agent, K. R., and C. V. Zegeer, "Effect of Pavement Texture on Traffic 

Noise," Transportation Research Record No. 602, Transportation Re- 
search Board, Washington (1977). 

3. American Concrete Paving Association, "Guideline for Texturing of 

Portland Cement Concrete Highway Pavements," Technical Bulletin 
No. 19, March 1975. 

4. Chamber lin, W, P., and D. E. Amsler, "Measuring Surface Texture of 

Concrete Pavements by the Sand-Patch Method," Report FHWA-NY-78-RR 
62, July 1978, 

5. Apostolos, J. A., R. N. Doty, B. G. Page, and G. B. Sherman, "California 

Skid Resistance Studies," California Division of Highways, Transpor- 
tation Laboratory, Report No. CA-DOT-TL-3126-9-74-10, February 1974. 

6. Balmer, Glenn G., "The Significance of Pavement Texture," Federal High- 

way Administration Report No. FHWA-RD-75-12, February 1975. 

7. Beaton, John L. , "Providing Skid Resistant Pavements," Transportation 

Research Record , No. 622, Transportation Research Board, Washington 
(1976). 

8. Brakey, B. A., "Design, Construction and Performance of Plant Mix Seals," 

Colorado Division of Highways, December 1972. 

9. Brunner, R. J., "Pavement Grooving," Final Report on Pennsylvania De- 

partment of Transportation Research Project, No. 69-1, January 1973. 

10. California Department of Transportation, "Safety Evaluation Instructions," 

Unpublished, January 1975. 

11. Cantor, C. C, "Automation of the Schonfeld Method of Highway Surface 

Texture Code Assignment," Vol. V, Specification of Single Line Instru- 
ment for Measuring Highway Surface Texture, Final Report to Maryland 
State Highway Administration, AW-076-111-046, December 1976. 



106 



'■ I 



12. Chamberlin, W. P., and D. E. Amsler, "Pilot Field Study of Concrete Pave- 

ment Texturing Methods," Highway Research Record , No. 389, Highway Re- 
search Board, Washington (1972). 

13. Colley, B. A., A. P. Christensen and W. S. Nowlen, "Factors Affecting 

Skid Resistance and Safety of Concrete Pavements," Highway Research 
Board, Special Report No. 101, Washington (1969). 

14. Cooper, D. R. C, "Measurement of Road Surface Texture by a Contactless 

Sensor," Transport and Road Research Laboratory Report 639. 

15. Cornfert, Guy-Michael, "Characterization of the Texture of Road Surfaces 

by Digital Processing of the ProfileometerData," Pennsylvania State 
University, Automotive Research Program, Report 579, November 1977. 

16. Dahir, S. H. , and J. J. Henry, "Alternatives for Optimization of Aggregate 

and Pavement Properties Related to Friction and Wear Resistance," 
Federal Highway Administration, Draft Final Report for Contract DOT- 
FH-11-8814, March 1977. 

17. Dahir, S. H. , and H. J. Lentz, "Laboratory Evaluation of Pavement Sur- 

face Texture Characteristics in Relation to Skid Resistance," Federal 
Highway Administration, Report No. FHWA-RD- 75-60, June 1972. 

18. Doty, R. N., "A Study of the Sand Patch and Outflow Meter Methods of 

Pavement Surface Texture Measurement," Paper presented at the ASTM 
1974 Annaul Meeting Symposium on Surface Texture and Standard Sur- 
faces, June 1974. 

19. Eaton, A. L. , "Tire Sound: It's Relationship with Surface Macrotexture 

and Skid Resistance," Pennsylvania State University, Automotive Re- 
search Program Report S73, August 1976. 

20. Farnsworth, E. E. , "Reduction of Wet Pavement Accidents on Metropolitan 

Los Angeles Freeways," California Department of Transportation (1973). 

21. Federal Highway Administration, "Texturing of Concrete Pavements and 

Bridge Decks," FHWA Notice No. 5080.59, September 1976. 

22. Callaway, B. M. , et al., "Effects of Pavement Surface Characteristics 

and Textures on Skid Resistance," Texas Transportation Institute Re- 
search Report 138-4, March 1971. 

23. Callaway, B. M. , and J. A. Epps, "Mixture Design Concepts, Laboratory 

Tests and Construction Guides for Open-Graded Bituminous Overlays," 
Texas Transportation Institute Research Report 36-lF, October 1974. 

107 



24. Gallaway, B. M. , and J. A. Epps, "Conventional Chip Seals as Corrective 

Measures for Improved Skid Resistance," Transportation Research Record , 
No. 523, Transportation Research Board, Washington (1974). 

25. Gallaway, B. M. , and J. G. Rose, "Macrotexture, Friction, Cross Slope, 

and Wheel Track Depression Measurements on 41 Typical Texas Highway 
Pavements," Texas Transportation Institute Research, Report 138-2, 
June 1970. 

26. Gallaway, B. M. , R. E. Schiller and J. G. Rose, "The Effects of Rain- 

fall Intensity, Pavement Cross-Slope, Surface Texture and Drainage 
Length on Pavement Water Depths," Texas Transportation Institute 
Research, Report 138-5, May 1971. 

27. Gallaway, Bob M. and Hisao Tomita, "Microtexture Measurements of Pave- 

ment Surfaces," Texas Transportation Institute Research, Report No. 
138-1, February 1970. 

28. Gee, S., and W. L. King, "Laser Measurement of Pavement Surface Tex- 

tures," Federal Highway Administration, Report No. FHWA-RD-74-12, 
March 1974. 

29. Gramling, W. L., and J. G. Hopkins III, "Aggregate-Skid Resistance Re- 

lationship as Applied to Pennsylvania Aggregates," Pennsylvania De- 
partment of Transportation Research, Report 65-4. 

30. Goodman, H. A., "Pavement Texture Measurement from a Moving Vehicle," 

Joint Road Friction Program, Report No. 19, Pennsylvania Department 
of Highways - Pennsylvania State University, March 1970. 

31. Gustafson, R. F., "Correlation of the Mean Hydraulic Radius with Skid 

Resistance: An Application of the Outflow Meter," Penn State Auto- 
motive Research Program, Report No. S53, Pennsylvania State University, 
August 1974. 

32. Hankins , K. D., "A Program for Reducing Skidding Accidents During Wet 

Weather," Transportation Research Record , No. 622, Transportation Re- 
search Board, Washington (1976). 

33. Hegmon, R. R. , and M. Mizoguchi, "Pavement Texture Measurement by the 

Sand Patch and Outflow Meter Methods," Automotive Safety Research Pro- 
gram, Report No. S40, Study No. 67-11, Pennsylvania State University, 
January 1970. 



108 



34. Henry, J. J., and R. R. Hegmon, "Pavement Texture Measurement and 

Evaluation," Special Technical Publication 583 , American Society 
for Testing and Materials, Philadelphia (1975). 

35. Highway Research Board, "Ontario Highway Department Develops a Photo- 

graphic Method for Measuring Pavement Surface Wear," Highway Research 
News , No. 36, Summer 1969. 

36. Holt, F., and G. Mosgrove, "Skid Resistance Photo-Interpretation 

Manual," Research and Development Division, Ontario Ministry of 
Transportation and Communications, November 1977. 

37. Howerter, E. D. , and T. J. Rudd , "Automation of the Schonfeld Method 

for Highway Surface Texture Classification," Paper presented at 
the 55th Annual Meeting of the Transportation Research Board, 
January 1976. 

38. Karr, J. I., and M. Guillory, "Evaluation of Minor Improvements -- 

Part 8: Grooved Pavements," California Division of Highways, 
December 1972. 

39. Karr, J. I., and M. Guillory, "Evaluation of Minor Improvements-- 

Part 9: Open-Graded Asphalt Concrete Overlays," California Depart- 
ment of Transportation, January 1972. 

40. Kay, R, A., and J. K. Stephens, "Porous Friction Courses and Roadway 

Surface Noise," Federal Highway Administration, Implementation 
Package 74-11, March 1975. 

41. Kearney, J., W. McAlpin, and W. C. Burnett, "Development of Specifi- 

cations for Skid-Resistant Asphalt Concrete," presented at the 51st 
Annual Meeting of the Highway Research Board, January 1972. 

42. Kummer, H. W. , and W. E. Meyer, "Tentative Skid Resistance Require- 

ments for Main Rural Highways," NCHRP Report 37 (1967). 

43. Lawther, J. M. , et al., "Characterization of Pavement Macrotexture by 

Profile Spectral Analysis," National Bureau of Standards, Report 
GCR75-35, June 1974. 

44. Ledbetter, W. B., et al,, "Evaluation of Full-Scale Experimental Con- 

crete Highway Finishes," Texas Transportation Institute Research, 
Report 141-4F, September 1974. 



109 



45. Lees, G., and I. E. D. Katekhda, "Prediction of Medium and High Speed 

Skid Resistance Values by Means of a Newly Developed Outflow Meter," 
Asphalt Paving Technology , Volume 43 (1974). 

46. Leland, T. J. N., et al., "Effects of Pavement Texture on Wet Runway 

Braking Performance," NASA Technical Note D-4324 (1968). 

47. Leu, M. C, and J. J. Henry, "Prediction of Skid Resistance as a 

Function of Speed from Pavement Texture," presented at the 57th 
Annual Meeting of the Transportation Research Board, January 1978. 

48. McLoed, N. W. , "A General Method of Design for Seal Coats and Surface 

Treatments," Association of Asphalt Paving Technologists, Volume 38 
(1969). 

49. Maupin, G. W. , Jr., "Virginia's Experience with Open-Graded Surface 

Mix," Transportation Research Record , No. 595, Transportation Re- 
search Board, Washington (1976). 

50. Mellott, D. B. , "Analysis of Surface Texture," Pennsylvania Department 

of Transportation, Research Project No. 70-25, October 1970. 

51. Mellott, D. B. , "Slurry Seal Surface Treatments," Final Report, 

Pennsylvania Department of Transportation Research Project No. 69- 
32, November 1976. 

52. Meyer, A. H. "Wearability of PC Concrete Pavement Finishes," Trans - 

portation Engineering Journal , August 1974. 

53. Mills, J. A., "A Skid Resistance Study of Four Western States," 

Highway Research Board, Special Report 101, Washington (1968). 

54. Mitrey, R. J., et al., "Effects of Selected Pavement Surface Textures 

on Tire Noise," New York State Department of Transportation, May 1975. 

55. Moore, D. F., "Prediction of Skid Resistance Gradient and Drainage 

Characteristics for Pavements," Highway Research Record , No. 131, 
Highway Research Board, Washington (1966). 

56. National Cooperative Highway Research Program, Synthesis of Highway 

Practice, No. 14, "Skid Resistance" (1972). 

57. National Cooperative Highway Research Program, Synthesis of Highway 

Practice No. 49, "Open-Graded Friction Courses" (1978). 



110 



58. Neal, W. , et al., "Portland Cement Concrete Pavement Texture Quality 

Investigation," California Department of Transportation, January 1975. 

59. Nixon, J. F., et al,, "Sprinkle Treatment for Skid Resistant Surfaces," 

Texas Department of State Highways and Public Transportation Research 
Report, 510-lF, April 1977. 

60. Orchard, D. F., et al., "A Quick Method of Measuring the Surface Texture 

of Aggregate," Australian Road Research Board, Fifth Conference, 
Canberra (1970). 

61. Ortgies, B. H. , "Interim Analysis and Design Guide for Designing Skid 

Resistant Pavement Surfaces," Iowa Department of Transportation, 
Revised, December 1975. 

62. Parker, Jerry L., "Tine Finishing of Concrete Pavement," Highway 

Focus , Vol. 8, No. 2, Federal Highway Administration, April 1976. 

63. Rasmussen, R. J., "Pavement Surface Texturing and Restoration for 

Highway Safety," presented at 53rd Annaul Meeting of the Highway 
Research Board, January 1974. 

64. Rice, J. M. , "Seasonal Variations in Pavement Skid Resistance," Public 

Roads , Vol. 40, No. 4, Federal Highway Administration, March 1977. 

65. Rizenbergs, R. L. , et al., "Skid Resistance of Pavements," Kentucky 

Department of Highways Research Report, No. 347, September 1972. 

66. Road Research Laboratory, "Instructions for Using the Portable Re- 

sistance Tester," Road Note No. 27, Department of Scientific and 
Industrial Research, London, England (1960). 

67. Rose, J. G., et al., "Macrotexture Measurements and Related Skid 

Resistance at Speeds from 20 to 60 Miles per Hour," Highway Research 
Record, No. 341, Highway Research Board, Washington (1970). 

68. Rose, J. G., et al., "Summary and Analysis of the Attributes of Methods 

of Surface Texture Measurement," Special Technical Publication , No. 53, 
American Society for Testing and Materials, Philadelphia, June 1972. 

69. Rudd, T. J., et al., "Computer Evaluation of Pavement Texture," Federal 

Highway Administration, Draft Final Report for Contract No. DOT-FH-11- 
9117, August 1977. 



Ill 



70. Runkle, S. N. , and D. C. Mahone, "Virginia's Wet Pavement Accident Re- 

duction Program," Transportation Research Record , No. 622, Transpor- 
tation Research Board, Washington (1976). 

71. Sabey, B. E. , and G. N. Lupton, "Measurement of Road Surface Texture 

Using Photogrammetry ," Road Research Laboratory, Report No. LR57 
(1967). 

72. Scarr, R. A., "Transverse Texturing with Metal Tines," Highway Focus « 

Vol. 8, No. 2, Federal Highway Administration, April 1976. 

73. Schonfeld, R. , "Photo-Interpretation of Pavement Skid Resistance," 

Highway Research Record , No. 311, Highway Research Board, Washington 
(1970). 

74. Schonfeld, R., "Photo-Interpretation of Pavement Skid Resistance in 

Practice," Transportation Research Record, No. 523, Transportation 
Research Board, Washington (1974). 

75. Schroeder, C. J., and J. V. Bergren, "An Evaluation of the Roto-Mill 

Profiles on Concrete Pavements in Iowa," Iowa Department of Trans- 
portation, November 1976. 

76. Schulze, K. H. , and L. Beckmann, "Friction Properties of Pavements at 

Different Speeds," Special Technical Publication 326, American Society 
for Testing and Materials, Philadelphia (1969). 

77. Smith, R. N. , and L. E. Elliot, "Evaluation of Minor Improvements — 

Part 8: Grooved Pavement, Supplemental Report," California Depart- 
ment of Transportation, September 1975. 

78. Smith, R. W. , J. M. Rice, and S. R. Spelman, "Design of Open-Graded 

Asphalt Friction Courses," Federal Highway Administration, Report 
No. FHWA-RD-74-2, January 1974. 

79. Smith, R. W. , "Determination of Asphalt Content for Open-Graded Asphalt 

Friction Courses Containing Lightweight Aggregate," Federal Highway 
Administration, supplemented to Report No. FHWA-RD-74-2, July 1975. 

80. Sprague, R. A., "Surface Roughness Measurement Using a White Light 

Speckle," Applied Optics . December 1972. 

81. Swertfager, W. E. , "Safety Grooving in Louisiana," Highway Focus , Vol. 8, 

No. 2, Federal Highway Administration, April 1976. 



112 



d 



82. Thornton, J. B., "Determination of Desireable Finish for Concrete Pave- 

ment," Georgia Department of Transportation Research, Project No. 7110, 
July 1974. 

83. Veres, R. F., et al., "Use of Tire Noise as a Measure of Pavement Macro- 

texture," Special Technical Publication 583 , American Society for 
Testing and Materials, Philadelphia (1975). 

84. Zipkes, E. , "The Influence of Grooving of Road Pavements on Accident 

Frequency," Transportation Research Record , No. 624, Transportation 
Research Board, Washington (1976). 

85. Zube, E., and J. Skog, "Skid Resistance of Screenings for Seal Coats," 

Highway Research Record , No. 236, Highway Research Board, Washington 
(1968). 

86. St. John, A. D., R. R. Blackburn, and D. W, Harwood, "Effectiveness 

and Alternative Skid Reduction Measures-- Voltmie II: Benefit-Cost 
Model," Federal Highway Administration, November 1978. 



113 



APPENDIX A 



FHWA DESIGN PROCEDURE FOR OPEN-GRADED ASPHALT MIXTURES 



114 



This appendix presents the design procedure recommended by the 
Federal Highway Administration for open-graded asphalt mixtures. The design 
procedure described is taken from the final draft of the National Cooperative 
Highway Research Program Synthesis of Highway Practice, entitled "Open-Graded 
Friction Courses." This procedure is essentially the same as that presented 
in a 1974 Federal Highway Administration report by Smith, Rice and Spelman,Z^/ 
but is updated to reflect the changes recommended in the supplement to that 
report prepared in 1975 by Smith. 79/ 



1.0 Material Requirements 

1.1 It is recommended that relatively pure carbonate aggregates or any 
aggregates known to polish be excluded from the coarse aggregate fraction 
(material retained on the No. 8 sieve). In addition, the coarse aggregate 
fraction should have at least 75 percent (by weight) of particles with at 
least two fractured faces and 90 percent with one or more fractured faces. 
The abrasion loss (AASHTO T 96) should not exceed 40 percent. 

1.2 Recommended Gradation for Open-Graded Asphalt Friction Course. 

Sieve Size^ / Percent Passing "./ 

1/2" 100 

3/8" 95-100 

#4 30-50 

#8 5-15 

#200 2-5 

_a/ U.S. Sieve Series. 

b_/ By volijme. (This is the same as by weight unless specific 
gravities of aggregates being combined are different.) 

1.3 The recommended grade of asphalt cement is AC-10, AC-20, or AR- 

40, AASHTO M 226. For AC-10 and AC-20, the M 226 Table 2 requirements should 
apply where such asphalt is available. AR-40 requirements are given in Table 
3 of M 226. 

2.0 Preliminary Data 

2.1 Test coarse and fine aggregates as received for the project for 
gradation unless otherwise provided. If mineral filler is submitted as a 

115 



separate item, it should also be tested for specification compliance. 
Analyze gradation results to determine if proportions of aggregates and 
batching operations proposed by the contractor will meet the job-mix formula 
and the specification limits of step 1.2. 

2.2 Determine bulk and apparent specific gravity for the coarse and 
fine aggregate fractions (retained and passing the No. 8 sieve) for each type 
of material submitted. Additional specific gravity tests are not warranted 
when the only distinction between aggregates is size of grading. Using the 
information verified in step 2.1, mathematically compute the bulk and apparent 
specific gravity for the coarse and fine aggregate fractions (retained and 
passing the No. 8 sieve) for the proposed job-mix gradation. 

2.3 Test the asphalt cement to be used for specification compliance 
(AASHTO M 226), viscosity-temperature data, and specific gravity at 77.0°F. 

3.0 Asphalt Content 

3.1 Determine the surface capacity of the aggregate fraction that is 
retained on a No. 4 sieve in accordance with the following procedure. 

Note: For highly absorptive aggregates, use the procedure described in 
s t ep 3.3. 

K^ is determined from the percent of SAE No. 10 oil retained, which 
represents the total effect of superficial area, the aggregate's absorptive 
properties, and surface roughness. 

3.1.1 Quarter out 105 g representative of the material passing 
the 3/8 in. sieve and retained on the No. 4 sieve. 

3.1.2 Dry sample on hot plate or in 230 + 9°F oven to constant 
weight and allow to cool. 

3.1.3 Weigh out 100.0 g and place in a metal funnel (top diam) 
3-1/2 in. , height 4-1/2 in. , orifice 1/2 in. , with a piece of No. 10 sieve 
soldered to the bottom of the opening) . 

3.1.4 Completely immerse specimen in SAE No. 10 lubricating oil 
for 5 min. 

3.1.5 Drain for 2 minutes. 

3.1.6 Place funnel containing sample in 140°F oven for 15 min of 
additional draining. 



116 



3.1.7 Pour sample from funnel into tared pan; cool, and reweigh 
sample to nearest 0.1 g. Subtract original weight and record difference as 
percent oil retained (based on 100 g of dry aggregate). 

(a) If specific gravity for the fraction is greater than 2-70 or 
less than 2.60, apply correction to oil retained, using formula at bottom 
of chart in Figure 7. 

(b) Start at the bottom of chart in Figure 7 with the corrected 
percent of oil retained; follow straightedge vertically upward to intersec- 
tion with the diagonal line; hold point, and follow the straightedge hori- 
zontally to the left. The value obtained is the surface constant for the 
retained fraction and is known as Kc. 

3.2 Determine the required asphalt content, which is based on weight 
of aggregate, from the following relationship. 

Percent asphalt = (2.0 Kr + 4.0) x liil 

(SG)ca 

Where K^ = surface constant 

(SG)ca = apparent specific gravity of coarse aggregate (3/8 in, 
to No. 4) 

3.3 For highly absorptive aggregates, use the following procedure for 
determining K^ and asphalt content. 

3.3.1 Follow the recommended design procedure from step 3.1 
through step 3.1.3. 

3.3.2 Follow the instructions in step 3.1.4, except immerse the 
specimen for 30 min. 

3.3.3 Follow the recommended procedure from step 3.1.5 through 
step 3.1.7. 

3.3.4 Pour the sample onto a clean, dry, absorptive cloth; obtain 
a saturated surface dry condition; pour sample from cloth into a tared pan; 
reweigh sample to nearest 0.1. Subtract original weight of aggregate and 
record difference as percent oil absorbed (based on 100 g of aggregate). 

3.3.5 Subtract the percent oil absorbed value (see 3.3.4 above) 
from the percent oil retained value (see 3.3.3 above), and obtain the per- 
cent (free) oil retained value. This value represents the percent oil re- 
tained value that would have been obtained had the aggregate been a nonab- 
sorptive type. The above technique allows one to evaluate the aggregate's 

117 



u 



I 



<. 
I— 
in 
2: 
o 

LU 

o 



oo 



3.0 
2.8 
2.6 
2.4 

2.2 
2.0 

1.8 
1.6 

1.4 

1.2 

1.0 



I I I I I I I 1 I 

/ 



1.5 



7 8 910 



PERCENT OIL RETAINED— CORRECTED FOR SPECIFIC GRAVITY 
OF AGGREGATE 

Material used: Aggregate--passing 3/8" sieve, retained 
on No. 4 sieve 
Oil— SAE No. 10 

Oil Retained Corrected (%) = Oil Retained (%) X 

apparent specific gravity of course aggregate 

2.65 



Figure 7 - Chart for Determining Surface Constant (K^,) 

of Coarse Aggregate 



118 



surface and shape characteristics without the overwhelming influence of a 
large quantity of absorbed oil. 

3.3.6 Follow the procedure recommended in steps 3.1.8 and 3.2. 
The only exception is that the percent (free) oil retained value is used 
(from step 3.3.5) to obtain K^- Thus, the asphalt quantity determined is 
the "effective" asphalt content. 

3.3.7 Follow the recommended procedure indicated through sections 
4 and 5. Because asphalt absorption is not presently included in the formula 
for the determination of fine aggregate content, it is particularly desirable 
that the effects of oil absorption in the K^ test be excluded in the case of 
the highly absorptive aggregate. 

3.3.8 Prepare a trial mixture using an asphalt content equal to 
or somewhat greater than (try to estimate amount that will be absorbed) the 
effective asphalt content determined in step 3.3.6 and also using the aggre- 
gate gradation as determined in step 3.3.7. Using a suitable technique, such 
as the test for maximum specific gravity of asphalt mixtures (AASHTO T 209), 
determine the actual quantity of asphalt absorbed (in percent, based on total 
weight of aggregate). 

3.3.9 Determine the total asphalt content of the subject mixture 
by adding the effective asphalt content (from step 3.3.6) to the absorbed 
asphalt content (from step 3.3.8). 

3.3.10 Follow the recommended procedure indicated in sections 6 
and 7, using the total asphalt content for all subsequent computations and 
trials (from step 3.3.9). 

4.0 Void Capacity of Coarse Aggregate 

4.1 Use the following procedure to determine the vibrated unit weight 
and void capacity of the coarse aggregate fraction (material retained on a 
No. 8 sieve) of the proposed job-mix gradation, 

4.1.1 Apparatus 

Rammer--A portable electromagnetic vibrating rammer as shown in 
Figure 8, having a frequency of 3,600 cycles per min , suitable for use with 
115-Vac. The rammer shall have a tamper foot and extension as shown in 
Figure 9, 

Mold--A solid-wall metal cylinder with a detachable metal base 
plate and a detachable metal guide- reference bar as shown in Figure 10, 



119 




Figure 8 - FHWA Vibratory Compaction Apparatus 

120 



•5/8"D 



12" 



7/8" 



nr 



tt 



^r\ 



^ 



TAMPER FOOT 
EXTENSION 



•15/16" HEX. SHAPE 



^ 



TAMPER FOOT 



«* ^ 



1-1/2" D 



5-7/8" D- 



Figure 9 - Tamper Foot and Extension 



121 




Figure 10 - Cylindrical Mold for Testing Granular Materials 



122 



Wooden Base — A plywood disc 15 in. in diam, 2 in. thick, with a 
cushion of rubber hose attached to the bottom. The disc shall be constructed 
so it can be firmly attached to the base plate of the compaction mold. 

Timer — A stopwatch or other timing device graduated in divisions 
of 1.0 sec and accurate to 1.0 sec, and capable of timing the unit for up 
to 2 min. An electric timing device or electrical circuits to start and 
stop the vibratory rammer may be used. 

Dial Indicator — A dial indicator graduated in 0.001- in. increments 
and having a travel range of 3.0 in. 

4.1.2 Sample: Select a 5-lb sample of the coarse aggregate frac- 
tion from the proposed job-mix formula as verified in step 2.1. 

4.1.3 Procedure 

(a) Pour the selected sample into the compaction mold and place 
the tamper foot on the sample. 

(b) Place the guide-reference bar over the shaft of the tamper foot 
and secure the bar to the mold with the thumb screws. 

(c) Place the vibratory rammer on the shaft of the tamper foot 
and vibrate for 15 sec. During the vibration period, the operator must exert 
just enough pressure on the hammer to maintain contact between the sample and 
the tamper foot, 

(d) Remove the vibratory rammer from the shaft of the tamper foot 
and brush any fines from the top of the tamper foot. Measure the thickness 
(t) of the compacted material to the nearest 0.001 in. 

Note: The thickness (t) of the compacted sample is determined by 
adding the dial reading, minus the thickness of the tamper foot, to 
the measured distance from the inside bottom of the mold and the 
end of the dial gauge when it is seated on the guide-reference bar 
with stem fully extended. 

4.1.4 Calculations 

Calculate the vibrated unit weight (X) as follows: 
X = 6912(w)/7T(d)2t(lb/ft^) 

Where w = wt of coarse aggregate fraction (lb) 
d = diam of compaction mold (in.) 

123 



If w = 5 lb and d = 6 in. : 

X = 305.58/t(lb/ft3) 

where t is in inches 
Determine the void capacity (VMA) as follows: 

VMA = 100(1 - X/Uc) (in percent) 

where U = bulk solid unit weight (lb/ft ) of the coarse 
aggregate fraction. U^ is calculated from bulk specific 
gravity, as determined in step 2.2, multiplied by 62.4 
lb/ft3. 

5.0 Optimum Content of Fine Aggregate 

5.1 Determine the optimum content of fine aggregate fraction using the 
following relationship: 



Y = 



[% VMA - V] - [(% AC) (X)/Ua] 



[(% VMA - V)/100] + [(X)/Uf] 



Where: Y = Percent passing the No. 8 sieve (by weight) 

X = Actual vibrated unit weight of coarse 

aggregate (retained on the No. 8 sieve) 

Uf = Theoretical bulk dry solid unit weight of 
fine aggregate (passing the No. 8 sieve) 

Ua = Unit weight of asphalt cement 

%AC = Percent asphalt by total weight of ag- 
gregate (2.0 Kj, + 4.0) 

V = Design percent air voids (15.0 percent) 

%VMA = Percent voids mineral aggregate of the 
coarse aggregate (retained on the No. 8 
sieve), which is 100 - (100)(X)/U^ 

Uc = Theoretical bulk dry solid unit weight of 
coarse aggregate (retained on the No. 8 
sieve) 

Note: X, U^, Uj,, and Uf are in pounds per cubic foot. 

124 



In the above relationship, asphalt absorption by aggregate has been 
assumed to negligible. Because asphalt absorption requirements are consid- 
ered in the test for Kc (see step 3.1), the estimated air voids of 15 per- 
cent in the mixture will actually be greater by an amount equivalent to the 
volume of asphalt absorbed, in percent. This condition, provides, if any- 
thing, an additional safety factor. 

As an alternative to the use of the mathematical relationship, one may 
use the design chart shown in Figure 11> provided that the assumptions used 
in designing the chart are satisfied; that is, the specific gravity values 
(bulk dry) for the coarse and fine aggregate fractions do not deviate beyond 
the limits of 2.600 to 2.700. 

If the value thus obtained for fine aggregate content is greater than 
15 percent, a value of 15.0 percent shall be used. 

5.2 Compare the optirnxom fine aggregate content (Y) determined in step 
5.1 to the amount passing the No. 8 sieve of the contractor's proposed job- 
mix formula. If these values differ by more than plus or minus 1 percentage 
point, reconstruct a revised 'j^r^-aS^^a^ted job-mix formula using the value 
determined for optimum fine aggregate content. Recompute the proportions 
of coarse and fine aggregates (as received) to meet the revised job-mix for- 
mula for submission to the contractor. 

Note: If the proposed and revised job-mix gradations are significantly 
different, it may be necessary to rerun portions of this procedure. 

6.0 Optimum Mixing Temperature 

6.1 Prepare a IjOOO-g sample of aggregate in the proportions determined 
in section 5. Mix this sample at the asphalt content determined in step 3.2 
at a temperature corresponding to an asphalt viscosity of 800 centistokes 
determined in step 2.3. When the mixture is completely coated, transfer it 
to a pyrex glass plate (8 to 9 in. diam) and spread the mixture with a minimum 
of manipulation. Return it to the oven at the mixing temperature. Observe 
the bottom of the plate after 15 and 60 min. A slight puddle at points of 
contact between aggregate and glass plate is suitable and desirable. Other- 
wise, repeat the test at a lower mixing temperature, or higher if necessary. 

Note: If asphalt drainage occurs at a mixing temperature that is too 
low to provide for adequate drying of the aggregate, an asphalt of a 
higher grade should be used. 

7.0 Resistance to Effects of Water 

7.1 Conduct the Immersion-Compression Test (AASHTO T 165 and T 167) on 
the designed mixture. Prepare samples at the optimiom mixing temperature 
determined in step 6.1. Use a molding pressure of 1,000 psi rather than the 
specified value of 3,000 psi. 

125 



UJ 

I— I 

I/) 

CO 

o 



</1 

a. 



UJ 

<£. 

UJ 
CO 

o 



o 



u. 
o 



o 

UJ 

a. 



UJ 

I— 
o 

UJ 

a: 

eg 
«j: 

UJ 




25 



30 



35 



40 



45 



PERCENT VOIDS (VMA) IN COARSE AGGREGATE (RETAINED ON NO. 8 SIEVE) 
Assumptions Used in Deriving Chart: 



U- = 165.4 Ib/ft^ 
Uf = 165.4 lb/ft:J 
Ug = 62.4 lb/ft>^ 
V = 15.0 percent 



(SG = 2.650) 
(SG = 2.650) 
(SG = 1.000) 



Figure 11 - Determination of Optimum Fine Aggregate Content 



126 



After a four-day immersion at 120 F, the index of retained strength shall 
not be less than 50 percent unless otherwise permitted. 

Note: Additives to promote adhesion that will provide adequate retained 
strength may be used when necessary. 

8.0 Design Calculations 

8.1 The following pages contain a convenient form for recording labora- 
tory test data and performing design calculations. 



127 



REPORT ON OPEN-GRADED ASPHALT FRICTION COURSE DESIGN 



1. AGGREGATES 



A. Proposed Proportions (by weight) 



B. Proposed Job-Mix Gradation 



Sieve 
size 


Sp 


ecification 
limits 


1/2 in. 






3/8 in. 


95-100 


No. 4 


30-50 


No. 8 


5-15 


No. 16 


- 


No. 200 


2-5 



Job-mix 
blend 



C. Specific Gravity — Unit Weight 



Apparent 
SG 



Bulk SG 
(dry basis) 



Bulk solid 
unit weight 
(lb/ft3) 



Coarse aggregate 
(retained on No. 8 sieve) 



(Up) 



Fine aggregate 
(passing No. 8 sieve) 

3/8 in. - No. 4 
Sieve fraction 



.(Uf) 



128 



D. Void Capacity of Coarse Aggregate 

Unit weight (vibrated, Ib/ft^) = (X) 

Voids mineral aggregate (%) = (VMA) 

E. K^ Determination 

Oil retention (g oil per 100 g aggregate) = 

Oil retention (corrected, 2.65 SG) = 

Kq (from chart) = 

2. ASPHALT 

A. Specific gravity — unit weight 

Specific gravity at 77 F (25 C) = 



Unit weight (Ib/ft^) = (Ua) 

B. Viscosity— Temperature 

Asphalt grade = 

Viscosity 
Temperature (°F) (centistokes) 

290 

275 

260 

245 

230 

215 



Target: ( - ) (700 - 900) 



129 



C. Asphalt Content (AC, %) 

Percent asphalt (aggregate basis) = 



(2K(, + 4) X 2.65 



apparent specific gravity of coarse aggregate 
(3/8 in. to No. 4 sieve) 



3, MIXTURE DESIGN 



A. 


Optimum Fine 


Aggregate 


Content (Y) 






Using: 
Where: 

Find: 


For mi 

X = 

Uf = 

Uc = 

Ua = 

Y = _ 


xla _ 




Chart 
_ Ib/ft^ 
_ lb/ft3 
. Ib/ft3 
„ Ib/ft^ 

(specs 


VMA = ; 








AC = ; 




V = ) 














% 


limit: 5 < Y < 



Remarks : 



B. Optimum Mixing Temperature 

Viscosity 
Temperature (°F) (centistokes) Drainage Use 



130 



C. Maximum Specific Gravity of Mixture (AASHTO T 209) 
Specific gravity (vacuum saturation) = ■ 

Unit weight (vacuum saturation) = Ib/ft^ 

D. Resistance to Effects of Water (AASHTO T 165 and T 167, 2000 psi) 
Air dry strength (psi) = 

Wet strength (psi) 
Retained strength 



Air voids 



Remarks ; 



'4 days at 120 F 
50% minimum 



4. DESIGN SUMMARY 



A. Aggregate Proportions (by weight) 



Bulk volume by 
dimensional measurement 



B. Job-Mix Gradation 



Percent passing 
Sieve size Job-mix blend 

1/2 in. 

3/8 in. 



No. 4 

No. 8 

No. 16 

No. 200 



131 



C. Asphalt Content 



D. Mixing Temperature 



E. Additives 



Aggregate basis ^'^"^ = 
Mixture basis 



Target value (°F) 



Range 



F. Recommendations 



Accepted Rejected 



132 



APPENDIX B 



TYPICAL SPECIFICATIONS FOR CONSTRUCTING 
AND MEASURING PAVEMENT GROOVES 



133 



This appendix presents a typical specification for constructing 
and measuring pavement grooves. The construction specification is used by 
the Louisiana Department of Highways and was presented in the April 1976 
issue of the Federal Highway Administration publication, Highway Focus . 
The specification for groove measurement constitutes Pennsylvania Depart- 
ment of Transportation Test Method No. 629 entitled, "Method of Test for 
Measuring Grooves in Concrete Pavements with a Modified Tire Tread Depth 
Gauge." The description of the measurement method includes two illustra- 
tive examples of its use. 



A. Construction Specification for Pavement Grooving 

ITEM S-1 GROOVING: The surface of the existing port land cement concrete 
pavement shall be grooved at the locations shown on the plans and grooving 
shall conform to the requirements of the plans and these specifications. 

Grooved areas shall begin and end at lines normal to the pavement center 
lane. The grooved area of each lane shall have a minimum width of 3 m (10 
ft) and shall be centered within the lane width. 

Grooving blades shall be 2,4 mm (0,095 in,) wide + 0,08 ram (+ 0.003 in.) 
and shall be spaced 19,1 mm (3/4 in,) on centers. The grooves shall be 
cut not less than 4,8 mm (3/16 in,) nor more than 7,9 ram (5/16 in,) deep. 

The actual grooved area of any selected 0.6 m (2 ft) by 30.5 m (100 ft) 
longitudinal area of pavement specified to be grooved shall be not less 
than 95 percent of the selected area. Any area within the selected area 
not grooved shall be due only to irregularities in the pavement surface 
and for no other reason. 

Residue from grooving operations -lihall not be permitted to flow across 
shoulders or lanes occupied by public traffic or to flow across shoulders 
or lanes occupied by public traffic or to flow into gutters or other drain- 
age facilities. Solid residue resulting from grooving operations shall be 
removed from pavement surfaces before such residue is blown by the action 
of traffic or wind. 

The contractor shall make every effort to insure that noise levels generated 
by the combined grooving operation shall not be in excess of those levels ) 
normally generated by existing truck traffic. 

Pavement grooving will be measured by the square meter. The quantity of pave- 
ment grooving to b^ paid for will be determined by multiplying the width of 
the grooved area by the total horizontal length of lane grooved. 



134 



The contract price per square yard for grooving existing concrete pavement 
shall include full compensation for furnishing all labor, materials, tools, 
equipment, and incidentals and for doing all work involved in grooving the 
existing concrete pavement, including removing residue, as shown on the 
plans, as specified in these special provisions, and as directed by the 
project engineer. Payment will be made under: 
Item S-1, Grooving, per square meter. 



B. Test Method for Measuring the Depth of Pavement Grooves 

1. Scope 

1.1 This method of test describes the procedure for sampling, 
preparing and measuring the depth of grooves in bridge decks, concrete pave- 
ments and ramps using a Modified Tire Tread Depth Gage. 

1.2 This method of test can measure the depth of grooves in 
concrete pavements produced by the following methods: Tine Finish, Broom 
Finish and Pavement Grooving, 

2. Apparatus 

2.1 Tire Tread Depth Gage - A gage, calibrated in increments 
of 0,5 mm (1/32 in.) and capable of measuring to a depth of 13 mm (1/2 in.) 
shall be used. The gage end shall be modified to a shape suitable for the 
measurement, 

2.2 Miscellaneous Equipment - Hand broom or brush, 0.3 m (12 in.) 
ruler, 31 m (100 ft.) tape measure and notebook. 

3. Sampling procedure for securing test area, 

3.1 The lot size for bridge decks shall be the length of the 
span by the width of the lane or lanes in one direction, 

3.2 The lot size for concrete pavements shall be a minimum 
of 4,181 sq m (5,000 sq yards) to a maximum of 8,362 sq m (10,000 sq yards) 
in one direction. If the contractor's production is below the minimimi, the 
lot size shall be the square yards of pavement placed, 

3.3 The lot size for ramps or separate lanes connecting with 
cross streets shall be the square yards of pavement placed in one d|rection. 

3.4 A lot shall consist of 5 approximately equal sublets. 

3.5 Within each sublot, one test area shall be randomly 
secured in accordance with PTM No. 1. 

135 



4. Preparing the test area. 

4.1 Brush all loose material from the area to be measured. 

5. Measuring the depth of the grooves. 

5.1 Measure ten grooves in a straight line perpendicular to 
the grooves, starting with the point that was randomly secured in Section 3.5. 

5.2 Place the Tire Tread Depth Gage on the groove to be mea- 
sured and firmly seat it to the surface. Make sure that the needle point 
will fall in the middle of the groove. 

5.3 Depress the needle point and determine the depth by read- 
ing the scale attached to the gage. 

5.4 Repeat the procedures described in Sections 5.2 and 5.3 
for the nine remaining grooves. 

6. Calculations 

6.1 Calculate the average groove depth for each of the 5 sub- 



lots. 



6.2 Calculate the average groove depth for the lot, 



7. Report 

7.1 The average groove depth for the lot shall be reported 
in increments of 0.5 ram (1/32 in,). 



ILLUSTRATIVE EXAMPLE NO. 1 













■^750'->" 


LOT 2 

3,360 m2 (10.000 /d^) 


SL 5 


SL 4 


LOT 1 

sl;3 


SL 2 


SL 1 

1,670 mi, 
(2,000 yd^) 



LOT 3 


LOT 4 



136 



Assume a contractor places 33,400 sq m (40,000 sq yards) of sep- 
arated highway consisting of reinforced cement concrete pavement 7,3 m (24 
ft) wide on each side of a traffic separator. 

In this case, the pavement can be divided into four lots. Each 
lot will have an area of 8,360 sq m (10,000 sq yards). 

Each lot must then be divided into five approximately equal sub- 
lots. Each sublet will have an area of 1,670 sq m (2,000 sq yards) (229 m 
(750 ft) by 7.3 m (24 ft)). 

Assume beginning station is 100 + 00, 

Use Table 2 from PTM No. 1 to obtain random decimal fractions in 
the X and Y columns. These values shall be multiplied by the length and 
width of the lanes of each sublet to obtain the coordinates of the sample 
location measured from the starting point of each sublet. 



Sublet #1 

Coordinate X = 0.47 x 229 m = 107 m or 0.47 x 750 ft = 352.5 ft 

Coordinate Y = R 0.20 x 7.3 m = R 1.5 m or R 0.20 x 24 ft = R 4.8 ft 

San5>le Location = 3ta. 100 + 00 plus 352.4 ft = Sta. 103 + 52.5 

Measure 4.8 ft from Rt. edge of lane 

Calculate the coordinates for the remaining sublets. 

Be sure to go through Table 2 before using the same ntmibers over. 



137 



ILLUSTRATIVE EXAMPLE NO. 2 




Assume a contractor places a ramp having an area of 1,400 square 



yards, 



In this case, the area of the ramp is the lot size. The lot shall 
begin where the uniform width starts and end at a point with a uniform width. 

The lot must then be divided into five approximately equal sublots. 
Each sublot will have an area of 234 m^ (280 sq yards) (64,0 m (210 ft) by 3.7 
m (12 ft)). 

Measurements for the sublots and X coordinates shall be made along 
the inner edge. 

The Y coordinate shall be measured on a line perpendicular to the 
sides of the ramp at the X coordinate point. 

See ILLUSTRATIVE EXAMPLE NO. 1 for example of how to obtain the 
coordinates of the sample location. 



OU.S. GOVERNMENT PRINTING OFFICE: 1979 633-809/475 1-3 



138 



FEDERALLY COORDINATED PROGRAM OF HIGHWAY 
RESEARCH AND DEVELOPMENT (FCP) 



The Offices of Research and Development of the 
Federal Highway Administration are responsible 
for a broad program of research with resources 
including its own staff, contract programs, and a 
Federal-Aid program which is conducted by or 
through the State highway departments and which 
also finances the National Cooperative Highway 
Research Program managed by the Transportation 
Research Board. The Federally Coordinated Pro- 
gram of Highway Research and Development 
(FCP) is a carefully selected group of projects 
aimed at urgent, national problems, which concen- 
trates these resources on these problems to obtain 
timely solutions. Virtually all of the available 
funds and staff resources are a part of the FCP. 
together with as much of the Federal-aid research 
funds of the States and the NCHRP resources as 
the States agree to devote to these projects."' 



FCP Category Descriptions 



n.Qc;«.n on/| Opera- 



ed with 
[ighway 
ety Act 
• design 
. and 
ulation 



and 



3. Environmental Considerations in High- 
way Design, Location, Construction, and 
Operation 

Environmental R&D is directed toward identify- 
ing and evaluating highway elements which 
affect the quality* of the human environment. 
The ultimate goals are reduction of adverse high- 
way and traffic impacts, and protection and 
enhancement of the environment. 

4. Improved Materials Utilization and Dura- 
bility 

Materials R&D is concerned with expanding the 
knowledge of materials properties and technology 
to fully utilize available naturally occurring 
materials, to develop extender or substitute ma- 
terials for materials in short supply, and to 
devise procedures for converting industrial and 
other wastes into useful highway products. 
These activities are all directed toward the com- 
mon goals of lowering the cost of highwav 
construction and extending the period of main- 
tenance-free operation. 

5. Improved Design to Reduce Costs, Extend 
Life Expectancy, and Insure Structural 
Safety 

Structural R&D is concerned with furthering the 
latest technological advances in structural de- 
signs, fabrication processes, and construction 
techniques, to provide safe, efficient highways 
at reasonable cost. 



ng the 
lys by 
5ns for 
r keep- 
better 
miques 
itment. 
affic. 



FCP is 
Service 
242057, 
oductoi-y 
Program 
lopnient, 
'. 20.500. 



6. Prototype Development and Implementa- 
tion of Research 

This category is concerned with developing and 
transferring research and technology into prac- 
tice, or, as it has been commonly identified, 
"technology transfer." 

7. Improved Technology for Highway Main- 
tenance 

Maintenance R&D objectives include the develop- 
ment and application of new technology to im- 
prove management, to augment the utilization 
of resources, and to increase operational efficiency 
and safety in the maintenance of highway 
facilities. 



ADDENDUM 

The charts on pages 100 and 101 are not intended for direct appli- 
cation. These charts were developed based upon a specific set 
of assumptions listed on page 70. The reader should carefully 
evaluate all assumptions as they pertain to his individual situ- 
ations. It should be also recognized that pavement friction 
demands are site specific and that generalized charts of this type 
are not. 



DDiaD<^MD